add sample module and tgnn trainer

a.py

+def foo():
+    a = 1
+    def fa():
+        print(a)
+        a+=1
+        print(a)
+    
+    def fb(a):
+        def apply():
+            print(a)
+        return apply
+    
+    fc = lambda: print(a)
+    return fa, fb(a), fc
+
+fa, fb, fc = foo()
+
+fa()
+fb()
+fc()

config/APAN.yml

+sampling:
+  - layer: 1
+    neighbor: 
+      - 10
+    strategy: 'recent'
+    prop_time: False
+    history: 1
+    duration: 0
+    num_thread: 32
+    no_neg: True
+memory: 
+  - type: 'node'
+    dim_time: 100
+    deliver_to: 'neighbors'
+    mail_combine: 'last'
+    memory_update: 'transformer'
+    attention_head: 2
+    mailbox_size: 10
+    combine_node_feature: False
+    dim_out: 100
+gnn:
+  - arch: 'identity'
+train:
+  - epoch: 100
+    batch_size: 600
+    lr: 0.0001
+    dropout: 0.1
+    att_dropout: 0.1
+    # all_on_gpu: True
\ No newline at end of file

config/DySAT.yml

+sampling:
+  - layer: 2
+    neighbor: 
+      - 10
+      - 10
+    strategy: 'uniform'
+    prop_time: True
+    history: 3
+    duration: 10000
+    num_thread: 32
+memory: 
+  - type: 'none'
+    dim_out: 0
+gnn:
+  - arch: 'transformer_attention'
+    layer: 2
+    att_head: 2
+    dim_time: 0
+    dim_out: 100
+    combine: 'rnn'
+train:
+  - epoch: 50
+    batch_size: 600
+    lr: 0.0001
+    dropout: 0.1
+    att_dropout: 0.1
+    all_on_gpu: True
\ No newline at end of file

config/JODIE.yml

+sampling:
+  - no_sample: True
+    history: 1
+memory: 
+  - type: 'node'
+    dim_time: 100
+    deliver_to: 'self'
+    mail_combine: 'last'
+    memory_update: 'rnn'
+    mailbox_size: 1
+    combine_node_feature: True
+    dim_out: 100
+gnn:
+  - arch: 'identity'
+    time_transform: 'JODIE'
+train:
+  - epoch: 100
+    batch_size: 600
+    lr: 0.0001
+    dropout: 0.1
+    all_on_gpu: True
\ No newline at end of file

config/TGAT.yml

+sampling:
+  - layer: 2
+    neighbor: 
+      - 10
+      - 10
+    strategy: 'uniform'
+    prop_time: False
+    history: 1
+    duration: 0
+    num_thread: 32
+memory: 
+  - type: 'none'
+    dim_out: 0
+gnn:
+  - arch: 'transformer_attention'
+    layer: 2
+    att_head: 2
+    dim_time: 100
+    dim_out: 100
+train:
+  - epoch: 10
+    batch_size: 600
+    lr: 0.0001
+    dropout: 0.1
+    att_dropout: 0.1
+    all_on_gpu: True
\ No newline at end of file

config/TGN.yml

+sampling:
+  - layer: 1
+    neighbor: 
+      - 10
+    strategy: 'recent'
+    prop_time: False
+    history: 1
+    duration: 0
+    num_thread: 32
+memory: 
+  - type: 'node'
+    dim_time: 100
+    deliver_to: 'self'
+    mail_combine: 'last'
+    memory_update: 'gru'
+    mailbox_size: 1
+    combine_node_feature: True
+    dim_out: 100
+gnn:
+  - arch: 'transformer_attention'
+    layer: 1
+    att_head: 2
+    dim_time: 100
+    dim_out: 100
+train:
+  - epoch: 10
+    #batch_size: 100
+    # reorder: 16
+    lr: 0.0001
+    dropout: 0.2
+    att_dropout: 0.2
+    all_on_gpu: True
\ No newline at end of file

config/readme.yml

+sampling:
+  - layer: <number of layers to sample>
+    neighbor: <a list of integers indicating how many neighbors are sampled in each layer>
+    strategy: <'recent' that samples most recent neighbors or 'uniform' that uniformly samples neighbors form the past>
+    prop_time: <False or True that specifies wherether to use the timestamp of the root nodes when sampling for their multi-hop neighbors>
+    history: <number of snapshots to sample on>
+    duration: <length in time of each snapshot, 0 for infinite length (used in non-snapshot-based methods)
+    num_thread: <number of threads of the sampler>
+memory: 
+  - type: <'node', we only support node memory now>
+    dim_time: <an integer, the dimension of the time embedding>
+    deliver_to: <'self' that delivers the mails only to involved nodes or 'neighbors' that deliver the mails to neighbors>
+    mail_combine: <'last' that use the latest latest mail as the input to the memory updater>
+    memory_update: <'gru' or 'rnn'>
+    mailbox_size: <an integer, the size of the mailbox for each node>
+    combine_node_feature: <False or True that specifies whether to combine node features (with the updated memory) as the input to the GNN.
+    dim_out: <an integer, the dimension of the output node memory>
+gnn:
+  - arch: <'transformer_attention' or 'identity' (no GNN)>
+    layer: <an integer, number of layers>
+    att_head: <an integer, number of attention heads>
+    dim_time: <an integer, the dimension of the time embedding>
+    dim_out: <an integer, the dimension of the output dynamic node embedding>
+train:
+  - epoch: <an integer, number of epochs to train>
+    batch_size: <an integer, the batch size (of edges); for multi-gpu training, this is the local batchsize>
+    reorder: <(optional) an integer that is divisible by batch size the specifies how many chunks per batch used in the random chunk scheduling>
+    lr: <floating point, learning rate>
+    dropout: <floating point, dropout>
+    att_dropout: <floating point, dropout for attention>
+    all_on_gpu: <False or True that decides if the node/edge features and node memory are completely stored on GPU>
\ No newline at end of file

data_maker.py

+
+import os
+import argparse
+import numpy as np
+import pandas as pd
+import torch
+from torch_geometric.data import Data
+
+from starrygl.sample.sample_core.base import NegativeSampling
+from starrygl.sample.part_utils.partition_tgnn import partition_save
+parser = argparse.ArgumentParser(
+    description="RPC Reinforcement Learning Example",
+    formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+)
+parser.add_argument('--data_name', default='WIKI', type=str, metavar='W',
+                    help='name of dataset')
+parser.add_argument('--num_neg_sample', default=1, type=int, metavar='W',
+                    help='number of negative samples')
+args = parser.parse_args()
+
+
+def load_feat(d, rand_de=0, rand_dn=0):
+    node_feats = None
+    if os.path.exists('DATA/{}/node_features.pt'.format(d)):
+        node_feats = torch.load('DATA/{}/node_features.pt'.format(d))
+        if node_feats.dtype == torch.bool:
+            node_feats = node_feats.type(torch.float32)
+    edge_feats = None
+    if os.path.exists('DATA/{}/edge_features.pt'.format(d)):
+        edge_feats = torch.load('DATA/{}/edge_features.pt'.format(d))
+        if edge_feats.dtype == torch.bool:
+            edge_feats = edge_feats.type(torch.float32)
+    if rand_de > 0:
+        if d == 'LASTFM':
+            edge_feats = torch.randn(1293103, rand_de)
+        elif d == 'MOOC':
+            edge_feats = torch.randn(411749, rand_de)
+    if rand_dn > 0:
+        if d == 'LASTFM':
+            node_feats = torch.randn(1980, rand_dn)
+        elif d == 'MOOC':
+            edge_feats = torch.randn(7144, rand_dn)
+    return node_feats, edge_feats
+
+
+data_name = args.data_name
+g = np.load('../tgl_main/DATA/'+data_name+'/ext_full.npz')
+df = pd.read_csv('../tgl_main/DATA/'+data_name+'/edges.csv')
+if os.path.exists('../tgl_main/DATA/'+data_name+'/node_features.pt'):
+    n_feat = torch.load('../tgl_main/DATA/'+data_name+'/node_features.pt')
+else:
+    n_feat = None
+if os.path.exists('../tgl_main/DATA/'+data_name+'/edge_features.pt'):
+    e_feat = torch.load('../tgl_main/DATA/'+data_name+'/edge_features.pt')
+else:
+    e_feat = None
+
+src = torch.from_numpy(np.array(df.src.values)).long()
+dst = torch.from_numpy(np.array(df.dst.values)).long()
+ts = torch.from_numpy(np.array(df.time.values)).long()
+neg_nums = args.num_neg_sample
+
+edge_index = torch.cat((src[np.newaxis, :], dst[np.newaxis, :]), 0)
+num_nodes = edge_index.view(-1).max().item()+1
+num_edges = edge_index.shape[1]
+print('the number of nodes in graph is {}, \
+      the number of edges in graph is {}'.format(num_nodes, num_edges))
+sample_graph = {}
+sample_src = torch.cat([src.view(-1, 1), dst.view(-1, 1)], dim=1)\
+    .reshape(1, -1)
+sample_dst = torch.cat([dst.view(-1, 1), src.view(-1, 1)], dim=1)\
+    .reshape(1, -1)
+sample_ts = torch.cat([ts.view(-1, 1), ts.view(-1, 1)], dim=1).reshape(-1)
+sample_eid = torch.arange(num_edges).view(-1, 1).repeat(1, 2).reshape(-1)
+sample_graph['edge_index'] = torch.cat([sample_src, sample_dst], dim=0)
+sample_graph['ts'] = sample_ts
+sample_graph['eids'] = sample_eid
+neg_sampler = NegativeSampling('triplet')
+neg_src = neg_sampler.sample(edge_index.shape[1]*neg_nums, num_nodes)
+neg_sample = neg_src.reshape(-1, neg_nums)
+
+
+edge_ts = torch.torch.from_numpy(np.array(ts)).float()
+data = Data()
+data.num_nodes = num_nodes
+data.num_edges = num_edges
+
+data.edge_index = edge_index
+data.edge_ts = edge_ts
+data.neg_sample = neg_sample
+
+if n_feat is not None:
+    data.x = n_feat
+if e_feat is not None:
+    data.edge_attr = e_feat
+
+data.train_mask = (torch.from_numpy(np.array(df.ext_roll.values)) == 0)
+data.test_mask = (torch.from_numpy(np.array(df.ext_roll.values)) == 1)
+data.val_mask = (torch.from_numpy(np.array(df.ext_roll.values)) == 2)
+sample_graph['train_mask'] = data.train_mask[sample_eid]
+sample_graph['test_mask'] = data.test_mask[sample_eid]
+sample_graph['val_mask'] = data.val_mask[sample_eid]
+data.sample_graph = sample_graph
+
+data.y = torch.zeros(edge_index.shape[1])
+edge_index_dict = {}
+edge_index_dict['edata'] = data.edge_index
+edge_index_dict['sample_data'] = data.sample_graph['edge_index']
+edge_index_dict['neg_data'] = torch.cat([neg_src.view(1, -1),
+                                         dst.view(-1, 1).repeat(1, neg_nums).
+                                         reshape(1, -1)], dim=0)
+data.edge_index_dict = edge_index_dict
+edge_weight_dict = {}
+edge_weight_dict['edata'] = 2*neg_nums
+edge_weight_dict['sample_data'] = 1*neg_nums
+edge_weight_dict['neg_data'] = 1
+partition_save('./dataset/here/'+data_name, data, 1, 'metis_for_tgnn',
+               edge_weight_dict=edge_weight_dict)
+partition_save('./dataset/here/'+data_name, data, 2, 'metis_for_tgnn',
+               edge_weight_dict=edge_weight_dict)
+#
+# partition_save('./dataset/here/'+data_name, data, 4, 'metis_for_tgnn',
+# edge_weight_dict=edge_weight_dict )
+# partition_save('./dataset/here'+data_name, data, 8, 'metis')
+# partition_save('./dataset/'+data_name, data, 12, 'metis')
+# partition_save('./dataset/'+data_name, data, 16, 'metis')
+

setup.sh

+#!/bin/sh
+#conda activate gnn
+cd ./starrygl/sample/sample_core
+if [ -f "setup.py" ]; then
+    rm -r build
+    rm sample_cores.cpython-*.so
+    python setup.py build_ext --inplace
+fi
+cd ../part_utils
+if [  -f "setup.py" ]; then
+    rm -r build
+    rm torch_utils.cpython-*.so
+    python setup.py build_ext --inplace
+fi
+cd ../../

starrygl/distributed/utils.py

@@ -93,6 +93,8 @@ class DistIndex:
     @property
     def dist(self) -> Tensor:
         return self.data
+    def to(self,device) -> Tensor:
+        return DistIndex(self.data.to(device))
 
 
 class DistributedTensor:
@@ -109,6 +111,21 @@ class DistributedTensor:
         self.num_nodes = DistInt(local_sizes)
         self.num_parts = DistInt([1] * len(self.rrefs))
 
+
+    @property
+    def dtype(self):
+        return self.accessor.data.dtype
+    
+    @property
+    def device(self):
+        return self.accessor.data.device
+    
+    def to(self,device):
+        return self.accessor.data.to(device)
+    
+    def __getitem__(self,index):
+        return self.accessor.data[index]
+    
     @property
     def ctx(self):
         return self.accessor.ctx

starrygl/module/layers.py

+from os.path import abspath, join, dirname
+import os
+import sys
+from os.path import abspath, join, dirname
+sys.path.insert(0, join(abspath(dirname(__file__))))
+import torch
+import dgl
+import math
+import numpy as np
+
+class TimeEncode(torch.nn.Module):
+
+    def __init__(self, dim):
+        super(TimeEncode, self).__init__()
+        self.dim = dim
+        self.w = torch.nn.Linear(1, dim)
+        self.w.weight = torch.nn.Parameter((torch.from_numpy(1 / 10 ** np.linspace(0, 9, dim, dtype=np.float32))).reshape(dim, -1))
+        self.w.bias = torch.nn.Parameter(torch.zeros(dim))
+
+    def forward(self, t):
+        output = torch.cos(self.w(t.float().reshape((-1, 1))))
+        return output
+
+class EdgePredictor(torch.nn.Module):
+
+    def __init__(self, dim_in):
+        super(EdgePredictor, self).__init__()
+        self.dim_in = dim_in
+        self.src_fc = torch.nn.Linear(dim_in, dim_in)
+        self.dst_fc = torch.nn.Linear(dim_in, dim_in)
+        self.out_fc = torch.nn.Linear(dim_in, 1)
+
+    def forward(self, h, neg_samples=1):
+        num_edge = h.shape[0] // (neg_samples + 2)
+        h_src =  self.src_fc(h[num_edge:2 * num_edge])#self.src_fc(h[:num_edge]) 
+        h_pos_dst = self.dst_fc(h[:num_edge]) #
+        h_neg_src = self.src_fc(h[2 * num_edge:])
+        h_pos_edge = torch.nn.functional.relu(h_src + h_pos_dst)
+        h_neg_edge = torch.nn.functional.relu(h_neg_src + h_pos_dst.tile(neg_samples, 1))
+        #h_neg_edge = torch.nn.functional.relu(h_neg_dst.tile(neg_samples, 1) + h_pos_dst)
+        #print(h_src,h_pos_dst,h_neg_dst)
+        return self.out_fc(h_pos_edge), self.out_fc(h_neg_edge)
+
+
+class TransfomerAttentionLayer(torch.nn.Module):
+
+    def __init__(self, dim_node_feat, dim_edge_feat, dim_time, num_head, dropout, att_dropout, dim_out, combined=False):
+        super(TransfomerAttentionLayer, self).__init__()
+        self.num_head = num_head
+        self.dim_node_feat = dim_node_feat
+        self.dim_edge_feat = dim_edge_feat
+        self.dim_time = dim_time
+        self.dim_out = dim_out
+        self.dropout = torch.nn.Dropout(dropout)
+        self.att_dropout = torch.nn.Dropout(att_dropout)
+        self.att_act = torch.nn.LeakyReLU(0.2)
+        self.combined = combined
+        if dim_time > 0:
+            self.time_enc = TimeEncode(dim_time)
+        if combined:
+            if dim_node_feat > 0:
+                self.w_q_n = torch.nn.Linear(dim_node_feat, dim_out)
+                self.w_k_n = torch.nn.Linear(dim_node_feat, dim_out)
+                self.w_v_n = torch.nn.Linear(dim_node_feat, dim_out)
+            if dim_edge_feat > 0:
+                self.w_k_e = torch.nn.Linear(dim_edge_feat, dim_out)
+                self.w_v_e = torch.nn.Linear(dim_edge_feat, dim_out)
+            if dim_time > 0:
+                self.w_q_t = torch.nn.Linear(dim_time, dim_out)
+                self.w_k_t = torch.nn.Linear(dim_time, dim_out)
+                self.w_v_t = torch.nn.Linear(dim_time, dim_out)
+        else:
+            if dim_node_feat + dim_time > 0:
+                self.w_q = torch.nn.Linear(dim_node_feat + dim_time, dim_out)
+            self.w_k = torch.nn.Linear(dim_node_feat + dim_edge_feat + dim_time, dim_out)
+            self.w_v = torch.nn.Linear(dim_node_feat + dim_edge_feat + dim_time, dim_out)
+        self.w_out = torch.nn.Linear(dim_node_feat + dim_out, dim_out)
+        self.layer_norm = torch.nn.LayerNorm(dim_out)
+
+    def forward(self, b):
+        assert(self.dim_time + self.dim_node_feat + self.dim_edge_feat > 0)
+        self.device = b.device
+        if b.num_edges() == 0:
+            return torch.zeros((b.num_dst_nodes(), self.dim_out), device=self.device)
+        if self.dim_time > 0:
+            time_feat = self.time_enc(b.edata['dt'])
+            zero_time_feat = self.time_enc(torch.zeros(b.num_dst_nodes(), dtype=torch.float32, device=self.device))
+        if self.combined:
+            Q = torch.zeros((b.num_edges(), self.dim_out), device=self.device)
+            K = torch.zeros((b.num_edges(), self.dim_out), device=self.device)
+            V = torch.zeros((b.num_edges(), self.dim_out), device=self.device)
+            if self.dim_node_feat > 0:
+                Q += self.w_q_n(b.srcdata['h'][:b.num_dst_nodes()])[b.edges()[1]]
+                K += self.w_k_n(b.srcdata['h'][b.num_dst_nodes():])[b.edges()[0] - b.num_dst_nodes()]
+                V += self.w_v_n(b.srcdata['h'][b.num_dst_nodes():])[b.edges()[0] - b.num_dst_nodes()]
+            if self.dim_edge_feat > 0:
+                K += self.w_k_e(b.edata['f'])
+                V += self.w_v_e(b.edata['f'])
+            if self.dim_time > 0:
+                Q += self.w_q_t(zero_time_feat)[b.edges()[1]]
+                K += self.w_k_t(time_feat)
+                V += self.w_v_t(time_feat)
+            Q = torch.reshape(Q, (Q.shape[0], self.num_head, -1))
+            K = torch.reshape(K, (K.shape[0], self.num_head, -1))
+            V = torch.reshape(V, (V.shape[0], self.num_head, -1))
+            att = dgl.ops.edge_softmax(b, self.att_act(torch.sum(Q*K, dim=2)))
+            att = self.att_dropout(att)
+            V = torch.reshape(V*att[:, :, None], (V.shape[0], -1))
+            b.edata['v'] = V
+            b.update_all(dgl.function.copy_edge('v', 'm'), dgl.function.sum('m', 'h'))
+        else:
+            if self.dim_time == 0 and self.dim_node_feat == 0:
+                Q = torch.ones((b.num_edges(), self.dim_out), device=self.device)
+                K = self.w_k(b.edata['f'])
+                V = self.w_v(b.edata['f'])
+            elif self.dim_time == 0 and self.dim_edge_feat == 0:
+                Q = self.w_q(b.srcdata['h'][:b.num_dst_nodes()])[b.edges()[1]]
+                K = self.w_k(b.srcdata['h'][b.edges()[0]])
+                V = self.w_v(b.srcdata['h'][b.edges()[0]])
+            elif self.dim_time == 0:
+                Q = self.w_q(b.srcdata['h'][:b.num_dst_nodes()])[b.edges()[1]]
+                K = self.w_k(torch.cat([b.srcdata['h'][b.edges()[0]], b.edata['f']], dim=1))
+                V = self.w_v(torch.cat([b.srcdata['h'][b.edges()[0]], b.edata['f']], dim=1))
+                #K = self.w_k(torch.cat([b.srcdata['h'][b.num_dst_nodes():], b.edata['f']], dim=1))
+                #V = self.w_v(torch.cat([b.srcdata['h'][b.num_dst_nodes():], b.edata['f']], dim=1))
+            elif self.dim_node_feat == 0:
+                Q = self.w_q(zero_time_feat)[b.edges()[1]]
+                K = self.w_k(torch.cat([b.edata['f'], time_feat], dim=1))
+                V = self.w_v(torch.cat([b.edata['f'], time_feat], dim=1))
+            elif self.dim_edge_feat == 0:
+                Q = self.w_q(torch.cat([b.srcdata['h'][:b.num_dst_nodes()], zero_time_feat], dim=1))[b.edges()[1]]
+                K = self.w_k(torch.cat([b.srcdata['h'][b.edges()[0]], time_feat], dim=1))
+                V = self.w_v(torch.cat([b.srcdata['h'][b.edges()[0]], time_feat], dim=1))
+                #K = self.w_k(torch.cat([b.srcdata['h'][b.num_dst_nodes():], time_feat], dim=1))
+                #V = self.w_v(torch.cat([b.srcdata['h'][b.num_dst_nodes():], time_feat], dim=1))
+            else:
+                Q = self.w_q(torch.cat([b.srcdata['h'][:b.num_dst_nodes()], zero_time_feat], dim=1))[b.edges()[1]]
+                K = self.w_k(torch.cat([b.srcdata['h'][b.edges()[0]], b.edata['f'], time_feat], dim=1))
+                V = self.w_v(torch.cat([b.srcdata['h'][b.edges()[0]], b.edata['f'], time_feat], dim=1))
+                #Q = self.w_q(torch.cat([b.srcdata['h'][:b.num_dst_nodes()], zero_time_feat], dim=1))[b.edges()[1]]
+                #K = self.w_k(torch.cat([b.srcdata['h'][b.num_dst_nodes():], b.edata['f'], time_feat], dim=1))
+                #V = self.w_v(torch.cat([b.srcdata['h'][b.num_dst_nodes():], b.edata['f'], time_feat], dim=1))
+            Q = torch.reshape(Q, (Q.shape[0], self.num_head, -1))
+            K = torch.reshape(K, (K.shape[0], self.num_head, -1))
+            V = torch.reshape(V, (V.shape[0], self.num_head, -1))
+            att = dgl.ops.edge_softmax(b, self.att_act(torch.sum(Q*K, dim=2)))
+            att = self.att_dropout(att)
+            V = torch.reshape(V*att[:, :, None], (V.shape[0], -1))
+            b.edata['v'] = V
+            b.update_all(dgl.function.copy_e('v', 'm'), dgl.function.sum('m', 'h'))
+            #b.srcdata['v'] = torch.cat([torch.zeros((b.num_dst_nodes(), V.shape[1]), device=torch.device('cuda:0')), V], dim=0)
+            #b.update_all(dgl.function.copy_u('v', 'm'), dgl.function.sum('m', 'h'))
+        if self.dim_node_feat != 0:
+            rst = torch.cat([b.dstdata['h'], b.srcdata['h'][:b.num_dst_nodes()]], dim=1)
+        else:
+            rst = b.dstdata['h']
+        rst = self.w_out(rst)
+        rst = torch.nn.functional.relu(self.dropout(rst))
+        return self.layer_norm(rst)
+
+class IdentityNormLayer(torch.nn.Module):
+
+    def __init__(self, dim_out):
+        super(IdentityNormLayer, self).__init__()
+        self.norm = torch.nn.LayerNorm(dim_out)
+
+    def forward(self, b):
+        return self.norm(b.srcdata['h'])
+
+class JODIETimeEmbedding(torch.nn.Module):
+
+    def __init__(self, dim_out):
+        super(JODIETimeEmbedding, self).__init__()
+        self.dim_out = dim_out
+
+        class NormalLinear(torch.nn.Linear):
+        # From Jodie code
+            def reset_parameters(self):
+                stdv = 1. / math.sqrt(self.weight.size(1))
+                self.weight.data.normal_(0, stdv)
+                if self.bias is not None:
+                    self.bias.data.normal_(0, stdv)
+
+        self.time_emb = NormalLinear(1, dim_out)
+    
+    def forward(self, h, mem_ts, ts):
+        time_diff = (ts - mem_ts) / (ts + 1)
+        rst = h * (1 + self.time_emb(time_diff.unsqueeze(1)))
+        return rst
+            
\ No newline at end of file

starrygl/module/memorys.py

+from os.path import abspath, join, dirname
+import os
+import sys
+from os.path import abspath, join, dirname
+sys.path.insert(0, join(abspath(dirname(__file__))))
+import torch
+import dgl
+from layers import TimeEncode
+from torch_scatter import scatter
+
+class MailBox():
+
+    def __init__(self, memory_param, num_nodes, dim_edge_feat, _node_memory=None, _node_memory_ts=None,_mailbox=None, _mailbox_ts=None, _next_mail_pos=None, _update_mail_pos=None):
+        self.memory_param = memory_param
+        self.dim_edge_feat = dim_edge_feat
+        if memory_param['type'] != 'node':
+            raise NotImplementedError
+        self.node_memory = torch.zeros((num_nodes, memory_param['dim_out']), dtype=torch.float32) if _node_memory is None else _node_memory
+        self.node_memory_ts = torch.zeros(num_nodes, dtype=torch.float32) if _node_memory_ts is None else _node_memory_ts
+        self.mailbox = torch.zeros((num_nodes, memory_param['mailbox_size'], 2 * memory_param['dim_out'] + dim_edge_feat), dtype=torch.float32) if _mailbox is None else _mailbox
+        self.mailbox_ts = torch.zeros((num_nodes, memory_param['mailbox_size']), dtype=torch.float32) if _mailbox_ts is None else _mailbox_ts
+        self.next_mail_pos = torch.zeros((num_nodes), dtype=torch.long) if _next_mail_pos is None else _next_mail_pos
+        self.update_mail_pos = _update_mail_pos
+        self.device = torch.device('cpu')
+        
+    def reset(self):
+        self.node_memory.fill_(0)
+        self.node_memory_ts.fill_(0)
+        self.mailbox.fill_(0)
+        self.mailbox_ts.fill_(0)
+        self.next_mail_pos.fill_(0)
+
+    def move_to_gpu(self):
+        self.node_memory = self.node_memory.cuda()
+        self.node_memory_ts = self.node_memory_ts.cuda()
+        self.mailbox = self.mailbox.cuda()
+        self.mailbox_ts = self.mailbox_ts.cuda()
+        self.next_mail_pos = self.next_mail_pos.cuda()
+        self.device = torch.device('cuda:0')
+
+    def allocate_pinned_memory_buffers(self, sample_param, batch_size):
+        limit = int(batch_size * 3.3)
+        if 'neighbor' in sample_param:
+            for i in sample_param['neighbor']:
+                limit *= i + 1
+        self.pinned_node_memory_buffs = list()
+        self.pinned_node_memory_ts_buffs = list()
+        self.pinned_mailbox_buffs = list()
+        self.pinned_mailbox_ts_buffs = list()
+        for _ in range(sample_param['history']):
+            self.pinned_node_memory_buffs.append(torch.zeros((limit, self.node_memory.shape[1]), pin_memory=True))
+            self.pinned_node_memory_ts_buffs.append(torch.zeros((limit,), pin_memory=True))
+            self.pinned_mailbox_buffs.append(torch.zeros((limit, self.mailbox.shape[1], self.mailbox.shape[2]), pin_memory=True))
+            self.pinned_mailbox_ts_buffs.append(torch.zeros((limit, self.mailbox_ts.shape[1]), pin_memory=True))
+
+    def prep_input_mails(self, mfg, global_id_list = None,use_pinned_buffers=False):
+        
+        if(global_id_list is not None):
+            idx = global_id_list[b.srcdata['ID']]
+        
+        else:
+            idx = b.srcdata['ID']
+        
+        if use_pinned_buffers:
+            idx = idx.cpu().long()
+            
+        for i, b in enumerate(mfg):
+            if use_pinned_buffers:
+                torch.index_select(self.node_memory, 0, idx, out=self.pinned_node_memory_buffs[i][:idx.shape[0]])
+                b.srcdata['mem'] = self.pinned_node_memory_buffs[i][:idx.shape[0]].cuda(non_blocking=True)
+                torch.index_select(self.node_memory_ts,0, idx, out=self.pinned_node_memory_ts_buffs[i][:idx.shape[0]])
+                b.srcdata['mem_ts'] = self.pinned_node_memory_ts_buffs[i][:idx.shape[0]].cuda(non_blocking=True)
+                torch.index_select(self.mailbox, 0, idx, out=self.pinned_mailbox_buffs[i][:idx.shape[0]])
+                b.srcdata['mem_input'] = self.pinned_mailbox_buffs[i][:idx.shape[0]].reshape(b.srcdata['ID'].shape[0], -1).cuda(non_blocking=True)
+                torch.index_select(self.mailbox_ts, 0, idx, out=self.pinned_mailbox_ts_buffs[i][:idx.shape[0]])
+                b.srcdata['mail_ts'] = self.pinned_mailbox_ts_buffs[i][:idx.shape[0]].cuda(non_blocking=True)
+            else:
+                b.srcdata['mem'] = self.node_memory[idx].cuda()
+                b.srcdata['mem_ts'] = self.node_memory_ts[idx].cuda()
+                b.srcdata['mem_input'] = self.mailbox[idx].cuda().reshape(b.srcdata['ID'].shape[0], -1)
+                b.srcdata['mail_ts'] = self.mailbox_ts[idx].cuda()
+
+    def update_memory(self, nid, memory, root_nodes, ts, global_node_list = None,neg_samples=1):
+        if nid is None:
+            return
+        num_true_src_dst = root_nodes.shape[0] // (neg_samples + 2) * 2
+        with torch.no_grad():
+            nid = nid[:num_true_src_dst].to(self.device)
+            memory = memory[:num_true_src_dst].to(self.device)
+            ts = ts[:num_true_src_dst].to(self.device)
+            self.node_memory[nid.long()] = memory
+            self.node_memory_ts[nid.long()] = ts
+
+    def update_mailbox(self, nid, memory, root_nodes, ts, edge_feats, block, global_node_list = None, neg_samples=1):
+        with torch.no_grad():
+            num_true_edges = root_nodes.shape[0] // (neg_samples + 2)
+            memory = memory.to(self.device)
+            if edge_feats is not None:
+                edge_feats = edge_feats.to(self.device)
+            if block is not None:
+                block = block.to(self.device)
+            # TGN/JODIE
+            if self.memory_param['deliver_to'] == 'self':
+                src = torch.from_numpy(root_nodes[:num_true_edges]).to(self.device)
+                dst = torch.from_numpy(root_nodes[num_true_edges:num_true_edges * 2]).to(self.device)
+                mem_src = memory[:num_true_edges]
+                mem_dst = memory[num_true_edges:num_true_edges * 2]
+                if self.dim_edge_feat > 0:
+                    src_mail = torch.cat([mem_src, mem_dst, edge_feats], dim=1)
+                    dst_mail = torch.cat([mem_dst, mem_src, edge_feats], dim=1)
+                else:
+                    src_mail = torch.cat([mem_src, mem_dst], dim=1)
+                    dst_mail = torch.cat([mem_dst, mem_src], dim=1)
+                mail = torch.cat([src_mail, dst_mail], dim=1).reshape(-1, src_mail.shape[1])
+                nid = torch.cat([src.unsqueeze(1), dst.unsqueeze(1)], dim=1).reshape(-1)
+                mail_ts = torch.from_numpy(ts[:num_true_edges * 2]).to(self.device)
+                if mail_ts.dtype == torch.float64:
+                    import pdb; pdb.set_trace()
+                # find unique nid to update mailbox
+                uni, inv = torch.unique(nid, return_inverse=True)
+                perm = torch.arange(inv.size(0), dtype=inv.dtype, device=inv.device)
+                perm = inv.new_empty(uni.size(0)).scatter_(0, inv, perm)
+                nid = nid[perm]
+                mail = mail[perm]
+                mail_ts = mail_ts[perm]
+                if self.memory_param['mail_combine'] == 'last':
+                    self.mailbox[nid.long(), self.next_mail_pos[nid.long()]] = mail
+                    self.mailbox_ts[nid.long(), self.next_mail_pos[nid.long()]] = mail_ts
+                    if self.memory_param['mailbox_size'] > 1:
+                        self.next_mail_pos[nid.long()] = torch.remainder(self.next_mail_pos[nid.long()] + 1, self.memory_param['mailbox_size'])
+            # APAN
+            elif self.memory_param['deliver_to'] == 'neighbors':
+                mem_src = memory[:num_true_edges]
+                mem_dst = memory[num_true_edges:num_true_edges * 2]
+                if self.dim_edge_feat > 0:
+                    src_mail = torch.cat([mem_src, mem_dst, edge_feats], dim=1)
+                    dst_mail = torch.cat([mem_dst, mem_src, edge_feats], dim=1)
+                else:
+                    src_mail = torch.cat([mem_src, mem_dst], dim=1)
+                    dst_mail = torch.cat([mem_dst, mem_src], dim=1)
+                mail = torch.cat([src_mail, dst_mail], dim=0)
+                mail = torch.cat([mail, mail[block.edges()[0].long()]], dim=0)
+                mail_ts = torch.from_numpy(ts[:num_true_edges * 2]).to(self.device)
+                mail_ts = torch.cat([mail_ts, mail_ts[block.edges()[0].long()]], dim=0)
+                if self.memory_param['mail_combine'] == 'mean':
+                    (nid, idx) = torch.unique(block.dstdata['ID'], return_inverse=True)
+                    mail = scatter(mail, idx, reduce='mean', dim=0)
+                    mail_ts = scatter(mail_ts, idx, reduce='mean')
+                    self.mailbox[nid.long(), self.next_mail_pos[nid.long()]] = mail
+                    self.mailbox_ts[nid.long(), self.next_mail_pos[nid.long()]] = mail_ts
+                elif self.memory_param['mail_combine'] == 'last':
+                    nid = block.dstdata['ID']
+                    # find unique nid to update mailbox
+                    uni, inv = torch.unique(nid, return_inverse=True)
+                    perm = torch.arange(inv.size(0), dtype=inv.dtype, device=inv.device)
+                    perm = inv.new_empty(uni.size(0)).scatter_(0, inv, perm)
+                    nid = nid[perm]
+                    mail = mail[perm]
+                    mail_ts = mail_ts[perm]
+                    self.mailbox[nid.long(), self.next_mail_pos[nid.long()]] = mail
+                    self.mailbox_ts[nid.long(), self.next_mail_pos[nid.long()]] = mail_ts
+                else:
+                    raise NotImplementedError
+                if self.memory_param['mailbox_size'] > 1:
+                    if self.update_mail_pos is None:
+                        self.next_mail_pos[nid.long()] = torch.remainder(self.next_mail_pos[nid.long()] + 1, self.memory_param['mailbox_size'])
+                    else:
+                        self.update_mail_pos[nid.long()] = 1
+            else:
+                raise NotImplementedError
+
+    def update_next_mail_pos(self):
+        if self.update_mail_pos is not None:
+            nid = torch.where(self.update_mail_pos == 1)[0]
+            self.next_mail_pos[nid] = torch.remainder(self.next_mail_pos[nid] + 1, self.memory_param['mailbox_size'])
+            self.update_mail_pos.fill_(0)
+
+class GRUMemeoryUpdater(torch.nn.Module):
+
+    def __init__(self, memory_param, dim_in, dim_hid, dim_time, dim_node_feat):
+        super(GRUMemeoryUpdater, self).__init__()
+        self.dim_hid = dim_hid
+        self.dim_node_feat = dim_node_feat
+        self.memory_param = memory_param
+        self.dim_time = dim_time
+        self.updater = torch.nn.GRUCell(dim_in + dim_time, dim_hid)
+        self.last_updated_memory = None
+        self.last_updated_ts = None
+        self.last_updated_nid = None
+        self.delta_memory = 0
+        if dim_time > 0:
+            self.time_enc = TimeEncode(dim_time)
+        if memory_param['combine_node_feature']:
+            if dim_node_feat > 0 and dim_node_feat != dim_hid:
+                self.node_feat_map = torch.nn.Linear(dim_node_feat, dim_hid)
+
+    def forward(self, mfg):
+        for b in mfg:
+            if self.dim_time > 0:
+                time_feat = self.time_enc(b.srcdata['ts'] - b.srcdata['mem_ts'])
+                b.srcdata['mem_input'] = torch.cat([b.srcdata['mem_input'], time_feat], dim=1)
+            updated_memory = self.updater(b.srcdata['mem_input'], b.srcdata['mem'])
+            self.last_updated_ts = b.srcdata['ts'].detach().clone()
+            self.last_updated_memory = updated_memory.detach().clone()
+            self.last_updated_nid = b.srcdata['ID'].detach().clone()
+            x1 = torch.sum(b.srcdata['mem']**2,dim = 1)
+            self.delta_memory = torch.sum((updated_memory - b.srcdata['mem'])**2,dim = 1)/torch.sum(b.srcdata['mem']**2,dim = 1)
+            #print(torch.dist(b.srcdata['mem'],updated_memory))
+            if self.memory_param['combine_node_feature']:
+                if self.dim_node_feat > 0:
+                    if self.dim_node_feat == self.dim_hid:
+                        b.srcdata['h'] += updated_memory
+                    else:
+                        b.srcdata['h'] = updated_memory + self.node_feat_map(b.srcdata['h'])
+                else:
+                    b.srcdata['h'] = updated_memory
+
+class RNNMemeoryUpdater(torch.nn.Module):
+
+    def __init__(self, memory_param, dim_in, dim_hid, dim_time, dim_node_feat):
+        super(RNNMemeoryUpdater, self).__init__()
+        self.dim_hid = dim_hid
+        self.dim_node_feat = dim_node_feat
+        self.memory_param = memory_param
+        self.dim_time = dim_time
+        self.updater = torch.nn.RNNCell(dim_in + dim_time, dim_hid)
+        self.last_updated_memory = None
+        self.last_updated_ts = None
+        self.last_updated_nid = None
+        self.delta_memory = 0
+        if dim_time > 0:
+            self.time_enc = TimeEncode(dim_time)
+        if memory_param['combine_node_feature']:
+            if dim_node_feat > 0 and dim_node_feat != dim_hid:
+                self.node_feat_map = torch.nn.Linear(dim_node_feat, dim_hid)
+
+    def forward(self, mfg):
+        for b in mfg:
+            if self.dim_time > 0:
+                #print(b.srcdata['ts'].shape,b.srcdata['mem_ts'].shape)
+                time_feat = self.time_enc(b.srcdata['ts'] - b.srcdata['mem_ts'])
+                b.srcdata['mem_input'] = torch.cat([b.srcdata['mem_input'], time_feat], dim=1)
+            updated_memory = self.updater(b.srcdata['mem_input'], b.srcdata['mem'])
+            self.last_updated_ts = b.srcdata['ts'].detach().clone()
+            self.last_updated_memory = updated_memory.detach().clone()
+            self.last_updated_nid = b.srcdata['ID'].detach().clone()
+            if self.memory_param['combine_node_feature']:
+                if self.dim_node_feat > 0:
+                    if self.dim_node_feat == self.dim_hid:
+                        b.srcdata['h'] += updated_memory
+                    else:
+                        b.srcdata['h'] = updated_memory + self.node_feat_map(b.srcdata['h'])
+                else:
+                    b.srcdata['h'] = updated_memory
+
+class TransformerMemoryUpdater(torch.nn.Module):
+
+    def __init__(self, memory_param, dim_in, dim_out, dim_time, train_param):
+        super(TransformerMemoryUpdater, self).__init__()
+        self.memory_param = memory_param
+        self.dim_time = dim_time
+        self.att_h = memory_param['attention_head']
+        if dim_time > 0:
+            self.time_enc = TimeEncode(dim_time)
+        self.w_q = torch.nn.Linear(dim_out, dim_out)
+        self.w_k = torch.nn.Linear(dim_in + dim_time, dim_out)
+        self.w_v = torch.nn.Linear(dim_in + dim_time, dim_out)
+        self.att_act = torch.nn.LeakyReLU(0.2)
+        self.layer_norm = torch.nn.LayerNorm(dim_out)
+        self.mlp = torch.nn.Linear(dim_out, dim_out)
+        self.dropout = torch.nn.Dropout(train_param['dropout'])
+        self.att_dropout = torch.nn.Dropout(train_param['att_dropout'])
+        self.last_updated_memory = None
+        self.last_updated_ts = None
+        self.last_updated_nid = None
+
+    def forward(self, mfg):
+        for b in mfg:
+            Q = self.w_q(b.srcdata['mem']).reshape((b.num_src_nodes(), self.att_h, -1))
+            mails = b.srcdata['mem_input'].reshape((b.num_src_nodes(), self.memory_param['mailbox_size'], -1))
+            if self.dim_time > 0:
+                time_feat = self.time_enc(b.srcdata['ts'][:, None] - b.srcdata['mail_ts']).reshape((b.num_src_nodes(), self.memory_param['mailbox_size'], -1))
+                mails = torch.cat([mails, time_feat], dim=2)
+            K = self.w_k(mails).reshape((b.num_src_nodes(), self.memory_param['mailbox_size'], self.att_h, -1))
+            V = self.w_v(mails).reshape((b.num_src_nodes(), self.memory_param['mailbox_size'], self.att_h, -1))
+            att = self.att_act((Q[:,None,:,:]*K).sum(dim=3))
+            att = torch.nn.functional.softmax(att, dim=1)
+            att = self.att_dropout(att)
+            rst = (att[:,:,:,None]*V).sum(dim=1)
+            rst = rst.reshape((rst.shape[0], -1))
+            rst += b.srcdata['mem']
+            rst = self.layer_norm(rst)
+            rst = self.mlp(rst)
+            rst = self.dropout(rst)
+            rst = torch.nn.functional.relu(rst)
+            b.srcdata['h'] = rst
+            self.last_updated_memory = rst.detach().clone()
+            self.last_updated_nid = b.srcdata['ID'].detach().clone()
+            self.last_updated_ts = b.srcdata['ts'].detach().clone()
+

starrygl/module/modules.py

+import torch
+import dgl
+from os.path import abspath, join, dirname
+import sys
+sys.path.insert(0, join(abspath(dirname(__file__))))
+from layers import *
+from memorys import *
+
+class GeneralModel(torch.nn.Module):
+
+    def __init__(self, dim_node, dim_edge, sample_param, memory_param, gnn_param, train_param, combined=False):
+        super(GeneralModel, self).__init__()
+        self.dim_node = dim_node
+        self.dim_node_input = dim_node
+        self.dim_edge = dim_edge
+        self.sample_param = sample_param
+        self.memory_param = memory_param
+        if not 'dim_out' in gnn_param:
+            gnn_param['dim_out'] = memory_param['dim_out']
+        self.gnn_param = gnn_param
+        self.train_param = train_param
+        if memory_param['type'] == 'node':
+            if memory_param['memory_update'] == 'gru':
+                self.memory_updater = RNNMemeoryUpdater(memory_param, 2 * memory_param['dim_out'] + dim_edge, memory_param['dim_out'], memory_param['dim_time'], dim_node)
+            elif memory_param['memory_update'] == 'rnn':
+                self.memory_updater = RNNMemeoryUpdater(memory_param, 2 * memory_param['dim_out'] + dim_edge, memory_param['dim_out'], memory_param['dim_time'], dim_node)
+            elif memory_param['memory_update'] == 'transformer':
+                self.memory_updater = TransformerMemoryUpdater(memory_param, 2 * memory_param['dim_out'] + dim_edge, memory_param['dim_out'], memory_param['dim_time'], train_param)
+            else:
+                raise NotImplementedError
+            self.dim_node_input = memory_param['dim_out']
+        self.layers = torch.nn.ModuleDict()
+        if gnn_param['arch'] == 'transformer_attention':
+            for h in range(sample_param['history']):
+                self.layers['l0h' + str(h)] = TransfomerAttentionLayer(self.dim_node_input, dim_edge, gnn_param['dim_time'], gnn_param['att_head'], train_param['dropout'], train_param['att_dropout'], gnn_param['dim_out'], combined=combined)
+            for l in range(1, gnn_param['layer']):
+                for h in range(sample_param['history']):
+                    self.layers['l' + str(l) + 'h' + str(h)] = TransfomerAttentionLayer(gnn_param['dim_out'], dim_edge, gnn_param['dim_time'], gnn_param['att_head'], train_param['dropout'], train_param['att_dropout'], gnn_param['dim_out'], combined=False)
+        elif gnn_param['arch'] == 'identity':
+            self.gnn_param['layer'] = 1
+            for h in range(sample_param['history']):
+                self.layers['l0h' + str(h)] = IdentityNormLayer(self.dim_node_input)
+                if 'time_transform' in gnn_param and gnn_param['time_transform'] == 'JODIE':
+                    self.layers['l0h' + str(h) + 't'] = JODIETimeEmbedding(gnn_param['dim_out'])
+        else:
+            raise NotImplementedError
+        self.edge_predictor = EdgePredictor(gnn_param['dim_out'])
+        if 'combine' in gnn_param and gnn_param['combine'] == 'rnn':
+            self.combiner = torch.nn.RNN(gnn_param['dim_out'], gnn_param['dim_out'])
+    
+                
+    def forward(self, mfgs, metadata = None,neg_samples=1):
+        if self.memory_param['type'] == 'node':
+            self.memory_updater(mfgs[0])
+        out = list()
+        for l in range(self.gnn_param['layer']):
+            for h in range(self.sample_param['history']):
+                rst = self.layers['l' + str(l) + 'h' + str(h)](mfgs[l][h])
+                if 'time_transform' in self.gnn_param and self.gnn_param['time_transform'] == 'JODIE':
+                    rst = self.layers['l0h' + str(h) + 't'](rst, mfgs[l][h].srcdata['mem_ts'], mfgs[l][h].srcdata['ts'])
+                if l != self.gnn_param['layer'] - 1:
+                    mfgs[l + 1][h].srcdata['h'] = rst
+                else:
+                    out.append(rst)
+        if self.sample_param['history'] == 1:
+            out = out[0]
+        else:
+            out = torch.stack(out, dim=0)
+            out = self.combiner(out)[0][-1, :, :]
+        #metadata需要在前面去重的时候记一下id
+        if metadata is not None:
+            #out = torch.cat((out[metadata['dst_pos_pos']],out[metadata['src_id_pos']],out[metadata['dst_neg_pos']]),0)
+            out = torch.cat((out[metadata['src_pos_index']],out[metadata['dst_pos_index']],out[metadata['src_neg_index']]),0)
+        return self.edge_predictor(out, neg_samples=neg_samples)
+
+    def get_emb(self, mfgs):
+        if self.memory_param['type'] == 'node':
+            self.memory_updater(mfgs[0])
+        out = list()
+        for l in range(self.gnn_param['layer']):
+            for h in range(self.sample_param['history']):
+                rst = self.layers['l' + str(l) + 'h' + str(h)](mfgs[l][h])
+                if 'time_transform' in self.gnn_param and self.gnn_param['time_transform'] == 'JODIE':
+                    rst = self.layers['l0h' + str(h) + 't'](rst, mfgs[l][h].srcdata['mem_ts'], mfgs[l][h].srcdata['ts'])
+                if l != self.gnn_param['layer'] - 1:
+                    mfgs[l + 1][h].srcdata['h'] = rst
+                else:
+                    out.append(rst)
+        if self.sample_param['history'] == 1:
+            out = out[0]
+        else:
+            out = torch.stack(out, dim=0)
+            out = self.combiner(out)[0][-1, :, :]
+        return out
+
+class NodeClassificationModel(torch.nn.Module):
+
+    def __init__(self, dim_in, dim_hid, num_class):
+        super(NodeClassificationModel, self).__init__()
+        self.fc1 = torch.nn.Linear(dim_in, dim_hid)
+        self.fc2 = torch.nn.Linear(dim_hid, num_class)
+
+    def forward(self, x):
+        x = self.fc1(x)
+        x = torch.nn.functional.relu(x)
+        x = self.fc2(x)
+        return x
\ No newline at end of file

starrygl/module/utils.py

+import yaml
+
+
+def parse_config(f):
+    conf = yaml.safe_load(open(f, 'r'))
+    sample_param = conf['sampling'][0]
+    memory_param = conf['memory'][0]
+    gnn_param = conf['gnn'][0]
+    train_param = conf['train'][0]
+    return sample_param, memory_param, gnn_param, train_param
\ No newline at end of file

starrygl/sample/batch_data.py

+from typing import List, Tuple
+import torch
+import torch.distributed as dist
+from starrygl.sample.graph_core import DataSet
+from starrygl.sample.graph_core import GraphData
+from starrygl.sample.sample_core.base import BaseSampler, NegativeSampling
+import dgl
+"""
+入参不变，出参变为：
+sample_from_nodes
+node: list[tensor,tensor, tensor...]
+eid: list[tensor,tensor, tensor...]
+src_index: list[tensor,tensor, tensor...]
+
+sample_from_edges：
+node
+eid: list[tensor,tensor, tensor...]
+src_index: list[tensor,tensor, tensor...]
+delta_ts: list[tensor,tensor, tensor...]
+metadata
+"""
+def prepare_input(node_feat, edge_feat, mem_embedding,mfgs,dist_nid,dist_eid):
+    for mfg in mfgs:
+        for i,b in enumerate(mfg):
+            e_idx = b.edata['ID']
+            idx = b.srcdata['ID']
+            b.edata['ID'] = dist_eid[e_idx]
+            b.srcdata['ID'] = dist_nid[idx]
+            if edge_feat is not None:
+                b.edata['f'] = edge_feat[e_idx]
+            if i == 0:
+                if node_feat is not None:
+                    b.srcdata['h'] = node_feat[idx]
+                if mem_embedding is not None:
+                    node_memory,node_memory_ts,mailbox,mailbox_ts = mem_embedding
+                    b.srcdata['mem'] = node_memory[idx]
+                    b.srcdata['mem_ts'] = node_memory_ts[idx]
+                    b.srcdata['mem_input'] = mailbox[idx].reshape(b.srcdata['ID'].shape[0], -1)
+                    b.srcdata['mail_ts'] = mailbox_ts[idx]
+                    #print(idx.shape[0],b.srcdata['mem_ts'].shape)
+        return mfgs
+
+def to_block(graph: GraphData, data, sample_out, mailbox = None,device = torch.device('cuda')):
+
+    if len(sample_out) > 1:
+        sample_out,metadata = sample_out
+    else:
+        metadata = None
+    eid = [ret.eid() for ret in sample_out]
+    eid_len = [e.shape[0] for e in eid ]
+    eid_mapper: torch.Tensor = graph.eids_mapper
+    nid_mapper: torch.Tensor = graph.nids_mapper
+    eid_tensor = torch.cat(eid,dim = 0).to(eid_mapper.device)
+    dist_eid = eid_mapper[eid_tensor].to(device)
+    dist_eid,eid_inv = dist_eid.unique(return_inverse=True)
+    src_node = graph.sample_graph['edge_index'][0,eid_tensor*2].to(graph.nids_mapper.device)
+    src_ts = None  
+    edge_feat = graph._get_edge_attr(dist_eid)
+    if metadata is None:
+        root_node = data.nodes.to(graph.nids_mapper.device)
+        root_len = [root_node.shape[0]]
+        if hasattr(data,'ts'):
+            src_ts = torch.cat([data.ts,
+                                graph.sample_graph['ts'][eid_tensor*2].to(device)])
+    elif 'seed' in metadata:
+        root_node = metadata.pop('seed').to(graph.nids_mapper.device)
+        root_len = root_node.shape[0]
+        if 'seed_ts' in metadata:
+            src_ts = torch.cat([metadata.pop('seed_ts').to(device),\
+                                graph.sample_graph['ts'][eid_tensor*2].to(device)])
+        for k in metadata:
+            metadata[k] = metadata[k].to(device)
+
+    nid_tensor = torch.cat([root_node,src_node],dim = 0)
+    dist_nid = nid_mapper[nid_tensor].to(device)
+    dist_nid,nid_inv = dist_nid.unique(return_inverse = True)
+    node_feat = graph._get_node_attr(dist_nid)
+    if mailbox is not None:
+        mem = mailbox._get_memory(dist_nid)
+    else:
+        mem = None
+
+    def build_block():
+        mfgs = list()
+        col = torch.arange(0,root_len,device = device)
+        col_len = 0
+        row_len = root_len
+        for r in range(len(eid_len)):
+            elen = eid_len[r]
+            row = torch.arange(row_len,row_len+elen,device = device)
+            b = dgl.create_block((row,col[sample_out[r].src_index().to(device)]),
+                                 num_src_nodes = row_len + elen,
+                                 num_dst_nodes = row_len,
+                                 device = device)
+            idx = nid_inv[0:row_len + elen]
+            e_idx = eid_inv[col_len:col_len+elen]
+            b.srcdata['ID'] = idx#dist_nid[idx]
+            if sample_out[r].delta_ts().shape[0] > 0:
+                b.edata['dt'] = sample_out[r].delta_ts().to(device)
+            if src_ts is not None:
+                b.srcdata['ts'] = src_ts[0:row_len + eid_len[r]]
+            b.edata['ID'] = e_idx#dist_eid[e_idx]
+            #if edge_feat is not None:
+            #    b.edata['f'] = edge_feat[e_idx]
+            #if r == len(eid_len)-1:
+            #    if node_feat is not None:
+            #        b.srcdata['h'] = node_feat[idx]
+            #    if mem_embedding is not None:
+            #        node_memory,node_memory_ts,mailbox,mailbox_ts = mem_embedding
+            #        b.srcdata['mem'] = node_memory[idx]
+            #        b.srcdata['mem_ts'] = node_memory_ts[idx]
+            #        b.srcdata['mem_input'] = mailbox[idx].cuda().reshape(b.srcdata['ID'].shape[0], -1)
+            #        b.srcdata['mail_ts'] = mailbox_ts[idx]
+            col = row
+            col_len += eid_len[r]
+            row_len += eid_len[r]
+            mfgs.append(b)
+        mfgs = list(map(list, zip(*[iter(mfgs)])))
+        mfgs.reverse()
+        return data,mfgs,metadata
+  
+    data,mfgs,metadata = build_block()
+    if dist.get_world_size() > 1:  
+        if(node_feat is None):
+            node_feat = torch.futures.Future()
+            node_feat.set_result(None)
+        if(edge_feat is None):
+            edge_feat = torch.futures.Future()
+            edge_feat.set_result(None) 
+        if(mem is None):
+            mem = torch.futures.Future()
+            mem.set_result(None)
+        def callback(fs,mfgs,dist_nid,dist_eid):
+            node_feat,edge_feat,mem_embedding = fs.value()
+            node_feat = node_feat.value()
+            edge_feat = edge_feat.value()
+            mem_embedding = mem_embedding.value()
+            return prepare_input(node_feat,edge_feat,mem_embedding,mfgs,dist_nid,dist_eid)
+        cal = lambda fut: callback(fs=fut,mfgs = mfgs,dist_nid = dist_nid,dist_eid =dist_eid)
+        return data,torch.futures.collect_all([node_feat,edge_feat,mem]).then(cal),metadata
+    else:
+        mfgs = prepare_input(node_feat,edge_feat,mem,mfgs,dist_nid,dist_eid)
+        #return build_block(node_feat,edge_feat,mem)#data,mfgs,metadata
+        return data,mfgs,metadata
+
+
+def graph_sample(graph, sampler:BaseSampler,
+                      sample_fn, data, 
+                      neg_sampling = None,
+                      mailbox = None,
+                      device = torch.device('cuda')):
+    out = sample_fn(sampler,data,neg_sampling)
+    return to_block(graph,data,out,mailbox,device)
+
+def sample_from_nodes(sampler:BaseSampler,  data:DataSet, **kwargs):
+    out = sampler.sample_from_nodes(nodes=data.nodes.reshape(-1))
+    #out.metadata = None
+    return out
+
+def sample_from_edges(sampler:BaseSampler,  
+                      data:DataSet, 
+                      neg_sampling:NegativeSampling = None):
+    edge_label = data.labels if hasattr(data,'labels') else None
+    out = sampler.sample_from_edges(edges = data.edges, 
+                                    neg_sampling=neg_sampling)
+    return out
+    
+
+def sample_from_temporal_nodes(sampler:BaseSampler,data:DataSet,
+                               **kwargs):
+    out = sampler.sample_from_nodes(nodes=data.nodes.reshape(-1),
+                                    ts = data.ts.reshape(-1))
+    #out.metadata = None
+    return out
+
+
+def sample_from_temporal_edges(sampler:BaseSampler, data:DataSet,
+                               neg_sampling: NegativeSampling = None):
+    edge_label = data.labels if hasattr(data,'labels') else None
+    out = sampler.sample_from_edges(edges=data.edges.to('cpu'),
+                                    ets=data.ts.to('cpu'),
+                                    neg_sampling = neg_sampling
+                                    )
+    return out
+
+class SAMPLE_TYPE:
+    SAMPLE_FROM_NODES = sample_from_nodes,
+    SAMPLE_FROM_EDGES = sample_from_edges,
+    SAMPLE_FROM_TEMPORAL_NODES = sample_from_temporal_nodes,
+    SAMPLE_FROM_TEMPORAL_EDGES = sample_from_temporal_edges
+    
\ No newline at end of file

starrygl/sample/data_loader.py

+from collections import deque
+from enum import Enum
+import queue
+import torch
+import sys
+from os.path import abspath, join, dirname
+import numpy as np
+from starrygl.sample.batch_data import graph_sample
+
+from starrygl.sample.sample_core.PreNegSampling import PreNegativeSampling
+
+sys.path.insert(0, join(abspath(dirname(__file__))))
+from typing import Deque, Optional
+import torch.distributed as dist
+from torch_geometric.data import Data
+import os.path as osp
+import math
+
+
+    
+
+class DistributedDataLoader:
+    ''' 
+     Args:
+            data_path: the path of loaded graph ,each part 0 of graph is saved on $path$/rank_0
+            num_replicas: the num of worker
+            
+
+    '''
+    def __init__(
+            self,
+            graph,
+            dataset = None,
+            sampler = None,
+            sampler_fn = None,
+            neg_sampler = None,
+            batch_size: Optional[int]=None,
+            drop_last = False,
+            device: torch.device  = torch.device('cuda'),
+            shuffle:bool = True,
+            chunk_size = None,
+            train = False,
+            queue_size = 10,
+            mailbox = None,
+            **kwargs
+    ):
+        assert sampler is not None
+        self.chunk_size = chunk_size
+        self.batch_size = batch_size
+        self.queue_size = queue_size
+        self.num_pending = 0
+        self.current_pos = 0
+        self.recv_idxs = 0
+        self.drop_last = drop_last
+        self.result_queue = deque(maxlen = self.queue_size)
+        self.shuffle = shuffle
+        self.is_closed = False
+        self.sampler = sampler
+        self.sampler_fn = sampler_fn
+        self.neg_sampler = neg_sampler
+        self.graph = graph
+        self.shuffle=shuffle
+        self.dataset = dataset
+        self.mailbox = mailbox
+        self.device =  device
+        if train is True:
+            self._get_expected_idx(self.dataset.len)
+        else:
+            self.expected_idx = int(math.ceil(self.dataset.len/self.batch_size))
+              
+    def __iter__(self):
+        if self.chunk_size is None:
+            if self.shuffle:
+                self.input_dataset = self.dataset.shuffle()
+            else:
+                self.input_dataset = self.dataset
+            self.recv_idxs = 0
+            self.current_pos = 0
+            self.num_pending = 0
+            self.submitted = 0
+        else:
+            self.input_dataset = self.dataset
+            self.recv_idxs = 0
+            self.num_pending = 0
+            self.submitted = 0
+            if dist.get_rank == 0:
+                self.current_pos = int(
+                    math.floor(
+                        np.random.uniform(0,self.batch_size/self.chunk_size)
+                    )*self.chunk_size
+                )
+            else:
+                self.current_pos = 0
+            current_pos = torch.tensor([self.current_pos],dtype = torch.long,device=self.device) 
+            dist.broadcast(current_pos, src = 0)
+            self.current_pos = int(current_pos.item())
+            self._get_expected_idx(self.dataset.len-self.current_pos)
+
+
+        if self.neg_sampler is not None \
+            and isinstance(self.neg_sampler,PreNegativeSampling):
+            self.neg_sampler.set_next_pos(self.current_pos)
+        return self
+
+    def _get_expected_idx(self,data_size):
+        world_size = dist.get_world_size()
+        self.expected_idx = data_size // self.batch_size if self.drop_last is True else int(math.ceil(data_size/self.batch_size))
+
+        if dist.get_world_size() > 1:
+            num_epochs = torch.tensor([self.expected_idx],dtype = torch.long,device=self.device) 
+            dist.all_reduce(num_epochs, op=dist.ReduceOp.MIN)
+            self.expected_idx = int(num_epochs.item())
+
+    def _next_data(self):   
+        if self.current_pos >= self.dataset.len:
+            return None
+
+        if self.current_pos + self.batch_size > self.input_dataset.len:
+            if self.drop_last:
+                return None
+            else:
+                next_data = self.input_dataset.get_next(
+                    slice(self.current_pos,None,None)
+                )
+                self.current_pos = 0
+        else:
+            next_data = self.input_dataset.get_next(
+                slice(self.current_pos,self.current_pos + self.batch_size,None)
+            )
+            self.current_pos += self.batch_size
+
+        return next_data
+    
+    def __next__(self):
+        if(dist.get_world_size() == 1):
+            if self.recv_idxs < self.expected_idx:
+                data = self._next_data()
+                batch_data = graph_sample(self.graph,
+                                          self.sampler,
+                                          self.sampler_fn,
+                                          data,self.neg_sampler,
+                                          self.mailbox,
+                                          self.device)
+                self.recv_idxs += 1
+                assert batch_data is not None
+                return batch_data
+            else :
+                raise StopIteration
+        else :
+            num_reqs = min(self.queue_size - self.num_pending,self.expected_idx - self.submitted)
+            for _ in range(num_reqs):
+                if len(self.result_queue)>0:
+                    result = self.result_queue[0]
+                    if result[1].done() == True:
+                        self.recv_idxs += 1
+                        self.num_pending -= 1
+                        self.result_queue.popleft()
+                        return result[0],result[1].value(),result[2]
+                else:
+                    next_data = self._next_data()
+                    assert next_data is not None
+                    batch_data = graph_sample(self.graph,
+                                          self.sampler,
+                                          self.sampler_fn,
+                                          next_data,self.neg_sampler,
+                                          self.mailbox,
+                                          self.device)
+                    self.result_queue.append(batch_data)
+                    self.submitted = self.submitted + 1
+                    self.num_pending = self.num_pending + 1
+                    
+            while(self.recv_idxs < self.expected_idx):
+                assert len(self.result_queue) > 0
+                result= self.result_queue[0]
+                if result[1].done() == True:
+                    self.recv_idxs += 1
+                    self.num_pending -= 1
+                    self.result_queue.popleft()
+                    return result[0],result[1].value(),result[2]
+                
+                
+            assert self.num_pending == 0
+            raise StopIteration
+                    
+
+        
+    
+
+
+    
+        
+        
+        

starrygl/sample/graph_core/__init__.py

+from starrygl.distributed.utils import DistIndex, DistributedTensor
+from starrygl.sample.graph_core.utils import build_mapper
+import os.path as osp
+import torch
+import torch.distributed as dist
+from torch_geometric.data import Data
+class GraphData():
+    def __init__(self, pdata, device = torch.device('cuda'), all_on_gpu = False):
+        self.device = device
+        self.ids = pdata.ids.to(device)
+        self.edge_index = pdata.edge_index.to(device)
+        if hasattr(pdata,'edge_ts'):
+            self.edge_ts = pdata.edge_ts.to(device).to(torch.float)
+        else:
+            self.edge_ts = None
+        self.sample_graph = pdata.sample_graph
+        self.nids_mapper = build_mapper(nids=pdata.ids.to(device)).data.to('cpu')
+        self.eids_mapper = build_mapper(nids=pdata.eids.to(device)).data.to('cpu')
+        if all_on_gpu:
+            self.nids_mapper = self.nids_mapper.to(device)
+            self.eids_mapper = self.eids_mapper.to(device)
+        self.num_nodes = self.nids_mapper.data.shape[0]
+        world_size = dist.get_world_size()
+        if hasattr(pdata,'x') and pdata.x is not None:
+            if world_size > 1:
+                self.x = DistributedTensor(pdata.x.to(self.device))
+            else:
+                self.x = pdata.x.to(device).to(torch.float)
+        else:
+            self.x = None
+        if hasattr(pdata,'edge_attr') and pdata.edge_attr is not None:
+            if world_size > 1:
+                self.edge_attr = DistributedTensor(pdata.edge_attr.to(self.device))
+            else:
+                self.edge_attr = pdata.edge_attr.to('cuda').to(torch.float)
+        else:
+            self.edge_attr = None
+
+    def _get_node_attr(self,ids):
+        if self.x is None:
+            return None
+        elif dist.get_world_size() == 1:
+            return self.x[ids]
+        else:
+            return self.x.index_select(ids)
+    def _get_edge_attr(self,ids):
+        if self.edge_attr is None:
+            return None
+        elif dist.get_world_size() == 1:
+            return self.edge_attr[ids.to('cpu')].to('cuda')
+        else:
+            return self.edge_attr.index_select(ids)
+
+class DataSet:
+    def __init__(self,nodes = None,
+                 edges = None,
+                 labels = None, 
+                 ts = None, 
+                 device = torch.device('cuda'),**kwargs):
+        if nodes is not None:
+            self.nodes = nodes.to(device)
+        if edges is not None:
+            self.edges = edges.to(device)
+        if ts is not None:
+            self.ts = ts.to(device)
+        if labels is not None:
+            self.labels = labels
+        self.len = self.nodes.shape[0] if nodes is not None else self.edges.shape[1] 
+        for k, v in kwargs.items():
+            assert isinstance(v,torch.Tensor) and v.shape[0]==self.len
+            setattr(self, k, v.to(device))
+
+    #@staticmethod
+    def get_next(self,indx):
+        nodes = self.nodes[indx] if hasattr(self,'nodes') else None
+        edges = self.edges[:,indx] if hasattr(self,'edges') else None
+        d = DataSet(nodes,edges)
+        for k,v in self.__dict__.items():
+            if k == 'edges' or k=='nodes' or k == 'len':
+                continue
+            else:
+                setattr(d,k,v[indx])
+        return d
+
+    #@staticmethod
+    def shuffle(self):
+        indx = torch.randperm(self.len)
+        nodes = self.nodes[indx] if hasattr(self,'nodes') else None
+        edges = self.edges[:,indx] if hasattr(self,'edges') else None
+        d = DataSet(nodes,edges)
+        for k,v in self.__dict__.items():
+            if k == 'edges' or k=='nodes' or k == 'len':
+                continue
+            else:
+                setattr(d,k,v[indx])
+        return d
+    
+class TemporalGraphData():
+    def __init__(self,pdata,device):
+        super(TemporalGraphData,self).__init__(pdata,device)
+    def _set_temporal_batch_cache(self,size,pin_size):
+        pass
+    def _load_feature_to_cuda(self,ids):
+        pass
+    
+
+
+
+class TemporalNeighborSampleGraph(GraphData):
+    def __init__(self, sample_graph=None, mode='full', eids_mapper=None):
+        self.edge_index = sample_graph['edge_index']
+        self.num_edges = self.edge_index.shape[1]
+        if 'ts' in sample_graph:
+            self.edge_ts = sample_graph['ts']
+        else:
+            self.edge_ts = None
+        self.eid = sample_graph['eids']
+        if mode == 'train':
+            mask = sample_graph['train_mask']
+        if mode == 'val':
+            mask = sample_graph['val_mask']
+        if mode == 'test':
+            mask = sample_graph['test_mask']
+        if mode != 'full':
+            self.edge_index = self.edge_index[:, mask]
+            self.edge_ts = self.edge_ts[mask]
+            self.eid = self.eid[mask]  
+
+
+        
+    
+
+

starrygl/sample/graph_core/utils.py

+from starrygl.distributed.context import DistributedContext
+from typing import *
+import torch.distributed as dist
+import torch
+
+from starrygl.distributed.utils import DistIndex
+def build_mapper(nids):
+    rank = dist.get_rank()
+    world_size = dist.get_world_size()
+    dst_len = nids.size(0)
+    ikw = dict(dtype=torch.long, device=nids.device)
+    num_nodes = torch.zeros(1, **ikw)
+    num_nodes[0] = dst_len
+    dist.all_reduce(num_nodes, op=dist.ReduceOp.SUM)
+    all_ids: List[torch.Tensor] = [None] * world_size
+    dist.all_gather_object(all_ids,nids)
+    part_mp = torch.empty(num_nodes,**ikw)
+    ind_mp = torch.empty(num_nodes,**ikw)
+    for i in range(world_size):
+        iid = all_ids[i]
+        part_mp[iid] = i
+        ind_mp[iid] = torch.arange(all_ids[i].shape[0],**ikw)
+    return DistIndex(ind_mp,part_mp)
+
+    
\ No newline at end of file

starrygl/sample/memory/shared_mailbox.py

+from typing import List
+import torch
+from torch.distributed import rpc
+import torch_scatter
+from starrygl.distributed.context import DistributedContext
+from starrygl.distributed.utils import DistIndex, DistributedTensor
+import torch.distributed as dist
+
+class SharedMailBox():
+    def __init__(self,
+                 num_nodes,
+                 memory_param,
+                 dim_edge_feat,
+                 device = torch.device('cuda')):
+        self.device = device
+        self.num_nodes = num_nodes
+        self.num_parts = dist.get_world_size()
+        if memory_param['type'] != 'node':
+            raise NotImplementedError
+        self.memory_param = memory_param
+        self.memory_size =  memory_param['dim_out']
+        node_memory = torch.zeros((
+            self.num_nodes, memory_param['dim_out']), 
+            dtype=torch.float32,device =self.device)
+        node_memory_ts = torch.zeros(self.num_nodes,
+                                      dtype=torch.float32,
+                                      device = self.device)
+        mailbox = torch.zeros(self.num_nodes, 
+                              memory_param['mailbox_size'], 
+                              2 * memory_param['dim_out'] + dim_edge_feat,
+                              device = self.device, dtype=torch.float32)
+        mailbox_ts = torch.zeros((self.num_nodes, 
+                                  memory_param['mailbox_size']), 
+                                 dtype=torch.float32,device  = self.device)
+        self.node_memory = DistributedTensor(node_memory)
+        self.node_memory_ts = DistributedTensor(node_memory_ts)
+        self.mailbox = DistributedTensor(mailbox)
+        self.mailbox_ts = DistributedTensor(mailbox_ts)
+        self.next_mail_pos = torch.zeros((self.num_nodes), 
+                                         dtype=torch.long,
+                                         device = self.device) 
+        self._ctx = DistributedContext.get_default_context()
+        self.rref = rpc.RRef(self)
+        self.rrefs = self._ctx.all_gather_remote_objects(self.rref)
+        self.partptr = torch.tensor([ ((i & 0xFFFF)<<48) for i in range(self.num_parts+1) ],device = device)
+
+
+    def reset(self):
+        self.node_memory.accessor.data.zero_()
+        self.node_memory_ts.accessor.data.zero_()
+        self.mailbox.accessor.data.zero_()
+        self.mailbox_ts.accessor.data.zero_()
+        self.next_mail_pos.zero_()
+    
+
+    def set_memory_local(self,index,source,source_ts,Reduce_Op = None):
+        if Reduce_Op == 'max':
+            unq_id,inv = index.unique(return_inverse = True)
+            max_ts,id =  torch_scatter.scatter_max(source_ts,inv,dim=0)
+            source_ts = max_ts
+            source = source[id]
+            index = unq_id
+        self.node_memory.accessor.data[index] = source
+        self.node_memory_ts.accessor.data[index] = source_ts
+
+
+    def set_mailbox_local(self,index,source,source_ts,Reduce_Op = None):
+        if Reduce_Op == 'max' and self.num_parts > 1:
+            unq_id,inv = index.unique(return_inverse = True)
+            max_ts,id =  torch_scatter.scatter_max(source_ts,inv,dim=0)
+            source_ts = max_ts
+            source = source[id]
+            index = unq_id
+        self.mailbox_ts.accessor.data[index, self.next_mail_pos[index]] = source_ts
+        self.mailbox.accessor.data[index, self.next_mail_pos[index]] = source
+        if self.memory_param['mailbox_size'] > 1:
+            self.next_mail_pos[index] = torch.remainder(
+                self.next_mail_pos[index] + 1, 
+                self.memory_param['mailbox_size'])
+            
+
+    def set_memory_async(self,index,source,source_ts):
+        dist_index = DistIndex(index)
+        part_idx = dist_index.part
+        index = dist_index.loc
+        futs: List[torch.futures.Future] = []
+        if self.num_parts == 1:
+            self.set_memory_local(index,source,source_ts)
+        for i in range(self.num_parts):
+            fut = self.ctx.remote_call(
+                SharedMailBox.set_memory_local,
+                self.rrefs[i],
+                index[part_idx == i], 
+                source[part_idx == i],
+                source_ts[part_idx == i])
+            futs.append(fut)
+        return torch.futures.collect_all(futs)
+
+
+    def add_to_mailbox_async(self,index,source,source_ts):
+        dist_index = DistIndex(index)
+        part_idx = dist_index.part
+        index = dist_index.loc
+        futs: List[torch.futures.Future] = []
+        if self.num_parts == 1:
+            self.set_mailbox_local(index,source,source_ts)
+        else:
+            for i in range(self.num_parts):
+                fut = self.ctx.remote_call(
+                    SharedMailBox.set_mailbox_local,
+                    self.rrefs[i],
+                    index[part_idx == i], 
+                    source[part_idx == i],
+                    source_ts[part_idx == i])
+                futs.append(fut)
+            return torch.futures.collect_all(futs)
+
+
+    def set_mailbox_all_to_all(self,index,memory,
+                               memory_ts,mail,mail_ts,
+                               reduce_Op = None):
+        #futs: List[torch.futures.Future] = []
+        if self.num_parts == 1:
+            dist_index = DistIndex(index)
+            part_idx = dist_index.part
+            index = dist_index.loc
+            self.set_mailbox_local(index,mail,mail_ts)
+            self.set_memory_local(index,memory,memory_ts)
+        else:
+            gather_len_list = torch.empty([self.num_parts],
+                                          dtype = int,
+                                          device = self.device)
+            indic = torch.searchsorted(index,self.partptr,right=False)
+            scatter_len_list = indic[1:] - indic[0:-1]
+            torch.distributed.all_to_all_single(gather_len_list,scatter_len_list)
+            input_split = scatter_len_list.tolist()
+            output_split = gather_len_list.tolist()
+            gather_id_list = torch.empty(
+                [gather_len_list.sum()],
+                dtype = torch.long,
+                device = self.device)
+            input_split = scatter_len_list.tolist()
+            output_split = gather_len_list.tolist()
+            torch.distributed.all_to_all_single(
+                gather_id_list,index,output_split_sizes=output_split, 
+                input_split_sizes=input_split)
+            index = gather_id_list
+            gather_memory = torch.empty(
+                [gather_len_list.sum(),memory.shape[1]],
+                dtype = memory.dtype,device = self.device)
+            gather_memory_ts = torch.empty(
+                [gather_len_list.sum()],
+                dtype = memory_ts.dtype,device = self.device)
+            gather_mail = torch.empty(
+                [gather_len_list.sum(),mail.shape[1]],
+                dtype = mail.dtype,device = self.device)
+            gather_mail_ts = torch.empty(
+                [gather_len_list.sum()],
+                dtype = mail_ts.dtype,device = self.device)
+            torch.distributed.all_to_all_single(
+                gather_memory,memory,
+                output_split_sizes=output_split, 
+                input_split_sizes=input_split)
+            torch.distributed.all_to_all_single(
+                gather_memory_ts,memory_ts,
+                output_split_sizes=output_split, 
+                input_split_sizes=input_split)
+            torch.distributed.all_to_all_single(
+                gather_mail,mail,
+                output_split_sizes=output_split, 
+                input_split_sizes=input_split)
+            torch.distributed.all_to_all_single(
+                gather_mail_ts,mail_ts,
+                output_split_sizes=output_split, 
+                input_split_sizes=input_split)
+            self.set_mailbox_local(DistIndex(index).loc,gather_mail,gather_mail_ts,Reduce_Op = reduce_Op)
+            self.set_memory_local(DistIndex(index).loc,gather_memory,gather_memory_ts, Reduce_Op = reduce_Op)
+
+
+    def get_update_mail(self,dist_indx_mapper,
+                 src,dst,ts,edge_feats,
+                 memory):
+        if edge_feats is not None:
+            edge_feats = edge_feats.to(self.device).to(self.mailbox.dtype)
+        src = src.to(self.device)
+        dst = dst.to(self.device)
+        index = torch.cat([src, dst]).reshape(-1)
+        index = dist_indx_mapper[index]
+
+        mem_src = memory[src]
+        mem_dst = memory[dst]
+        if edge_feats is not None:
+            src_mail = torch.cat([mem_src, mem_dst, edge_feats], dim=1)
+            dst_mail = torch.cat([mem_dst, mem_src, edge_feats], dim=1)
+        else:
+            src_mail = torch.cat([mem_src, mem_dst], dim=1)
+            dst_mail = torch.cat([mem_dst, mem_src], dim=1)
+        mail = torch.cat([src_mail, dst_mail], dim=1).reshape(-1, src_mail.shape[1])
+        mail_ts = torch.cat((ts,ts),-1).to(self.device).to(self.mailbox_ts.dtype)
+        unq_index,inv = torch.unique(index,return_inverse = True)
+        max_ts,idx = torch_scatter.scatter_max(mail_ts,inv,0)
+        mail_ts = max_ts
+        mail = mail[idx]
+        index = unq_index
+        return index,mail,mail_ts
+    
+
+    def get_update_memory(self,index,memory,memory_ts):
+        unq_index,inv = torch.unique(index,return_inverse = True)
+        max_ts,idx = torch_scatter.scatter_max(memory_ts,inv,0)
+        ts = max_ts
+        memory = memory[idx]
+        index = unq_index
+        return index,memory,ts
+    
+
+    def _get_memory(self,index):
+        if self.num_parts == 1:
+            return self.node_memory.accessor.data[index],\
+                    self.node_memory_ts.accessor.data[index],\
+                        self.mailbox.accessor.data[index],\
+                            self.mailbox_ts.accessor.data[index]
+        else:
+            memory = self.node_memory.index_select(index)
+            memory_ts = self.node_memory_ts.index_select(index)
+            mail = self.mailbox.index_select(index)
+            mail_ts = self.mailbox_ts.index_select(index)
+            def callback(fs):
+                memory,memory_ts,mail,mail_ts = fs.value()
+                memory = memory.value()
+                memory_ts = memory_ts.value()
+                mail = mail.value()
+                mail_ts = mail_ts.value()
+                #print(memory.shape[0])
+                return memory,memory_ts,mail,mail_ts
+            return torch.futures.collect_all([memory,memory_ts,mail,mail_ts]).then(callback)
+        
+
+

starrygl/sample/part_utils/partition_tgnn.py

+import parser
+from torch_sparse import SparseTensor
+from torch_geometric.data import Data
+from torch_geometric.utils import degree
+from .torch_utils import get_norm_temporal
+import os.path as osp
+import os
+import shutil
+import torch
+import torch.utils.data
+import metis
+import networkx as nx
+import torch.distributed as dist
+
+
+def partition_load(root: str, algo: str = "metis") -> Data:
+    rank = dist.get_rank()
+    world_size = dist.get_world_size()
+    fn = osp.join(root, f"{algo}_{world_size}", f"{rank:03d}")
+    return torch.load(fn)
+
+
+def partition_save(root: str, data: Data, num_parts: int,
+                   algo: str = "metis",
+                   edge_weight_dict=None):
+    root = osp.abspath(root)
+    if osp.exists(root) and not osp.isdir(root):
+        raise ValueError(f"path '{root}' should be a directory")
+    
+    path = osp.join(root, f"{algo}_{num_parts}")
+    if osp.exists(path) and not osp.isdir(path):
+        raise ValueError(f"path '{path}' should be a directory")
+    
+    if osp.exists(path) and os.listdir(path):
+        print(f"directory '{path}' not empty and cleared")
+        for p in os.listdir(path):
+            p = osp.join(path, p)
+            if osp.isdir(p):
+                shutil.rmtree(osp.join(path, p))
+            else:
+                os.remove(p)
+                
+    if not osp.exists(path):
+        print(f"creating directory '{path}'")
+        os.makedirs(path)
+    if algo == 'metis_for_tgnn':
+        for i, pdata in enumerate(partition_data_for_tgnn(
+                data, num_parts, algo, verbose=True,
+                edge_weight_dict=edge_weight_dict)):
+            print(f"saving partition data: {i+1}/{num_parts}")
+            fn = osp.join(path, f"{i:03d}")
+            torch.save(pdata, fn)
+    else:
+        for i, pdata in enumerate(partition_data_for_gnn(data, num_parts,
+                                                         algo, verbose=True)):
+            print(f"saving partition data: {i+1}/{num_parts}")
+            fn = osp.join(path, f"{i:03d}")
+            torch.save(pdata, fn)
+
+
+def partition_data_for_gnn(data: Data, num_parts: int, 
+                           algo: str, verbose: bool = False):
+    if algo == "metis":
+        part_fn = metis_partition
+
+    else:
+        raise ValueError(f"invalid algorithm: {algo}")
+    
+    num_nodes = data.num_nodes
+    num_edges = data.num_edges
+    edge_index = data.edge_index
+    
+    if verbose:
+        print(f"running partition algorithm: {algo}")
+    node_parts, edge_parts = part_fn(edge_index, num_nodes, num_parts)
+    
+    if verbose:
+        print("computing GCN normalized factor")
+    gcn_norm = compute_gcn_norm(edge_index, num_nodes)
+    
+    if data.y.dtype == torch.long:
+        if verbose:
+            print("compute num_classes")
+        num_classes = data.y.max().item() + 1
+    else:
+        num_classes = None
+        eids = torch.zeros(num_edges, dtype=torch.long)
+    len = 0
+    edgeptr = torch.zeros(num_parts+1, dtype=eids.dtype)
+    for i in range(num_parts):
+        epart_i = torch.where(edge_parts == i)[0]
+        eids[epart_i] = torch.arange(epart_i.shape[0]) + len
+        len += epart_i.shape[0]
+        edgeptr[i+1] = len
+    data.eids = eids
+    data.sample_graph.sample_eids = eids[data.sample_graph.sample_eid]
+    nids = torch.zeros(num_nodes, dtype=torch.long)
+    len = 0
+    partptr = torch.zeros(num_parts+1, dtype=nids.dtype)
+    for i in range(num_parts):
+        npart_i = torch.where(node_parts == i)[0]
+        nids[npart_i] = torch.arange(npart_i.shape[0]) + len
+        len += npart_i.shape[0]
+        partptr[i+1] = len
+    data.edge_index = nids[data.edge_index]
+    data.sample_graph.edge_index = nids[data.sample_graph.edge_index]
+    for i in range(num_parts):
+        npart_i = torch.where(node_parts == i)[0]
+        epart_i = torch.where(edge_parts == i)[0]
+        
+        npart = npart_i
+        epart = edge_index[:, epart_i]
+        
+        pdata = {
+            "ids": npart,
+            "edge_index": epart,
+            "gcn_norm": gcn_norm[epart_i],
+            "sample_graph": data.sample_graph,
+            "partptr": partptr,
+            "edgeptr": edgeptr
+        }
+
+        if num_classes is not None:
+            pdata["num_classes"] = num_classes
+    
+        for key, val in data:   
+            if key == "edge_index" or key == "sample_graph":
+                continue
+            if isinstance(val, torch.Tensor):
+                if val.size(0) == num_nodes:
+                    pdata[key] = val[npart_i]
+                elif val.size(0) == num_edges:
+                    pdata[key] = val[epart_i]
+                # else:
+                #     pdata[key] = val
+            elif isinstance(val, SparseTensor):
+                pass
+            else:
+                pdata[key] = val
+        
+        pdata = Data(**pdata)
+        yield pdata
+
+
+def _nopart(edge_index: torch.LongTensor, num_nodes: int):
+    node_parts = torch.zeros(num_nodes, dtype=torch.long)
+    if isinstance(edge_index, torch.Tensor):
+        edge_parts = torch.zeros(edge_index.size(1), dtype=torch.long)
+        return node_parts, edge_parts
+    return node_parts
+
+
+def metis_for_tgnn(edge_index_dict: dict,
+                   num_nodes: int,
+                   num_parts: int,
+                   edge_weight_dict=None):
+    if num_parts <= 1:
+        return _nopart(edge_index_dict, num_nodes)
+    G = nx.Graph()
+    G.add_nodes_from(torch.arange(0, num_nodes).tolist())
+    value, counts = torch.unique(edge_index_dict['edata'][1, :].view(-1),
+                                 return_counts=True)
+    nodes = torch.tensor(list(G.adj.keys()))
+    for i in range(value.shape[0]):
+        if (value[i].item() in G.nodes):
+            G.nodes[int(value[i].item())]['weight'] = counts[i]
+            G.nodes[int(value[i].item())]['ones'] = 1
+    G.graph['node_weight_attr'] = ['weight', 'ones']
+    for i, key in enumerate(edge_index_dict):
+        v = edge_index_dict[key]
+        edges = torch.cat((v, (torch.ones(v.shape[1], dtype=torch.long) *
+                               edge_weight_dict[key]).unsqueeze(0)), dim=0)
+        # w = edges.T
+        G.add_weighted_edges_from((edges.T).tolist())
+    G.graph['edge_weight_attr'] = 'weight'
+    cuts, part = metis.part_graph(G, num_parts)
+    node_parts = torch.zeros(num_nodes, dtype=torch.long)
+    node_parts[nodes] = torch.tensor(part)
+    return node_parts
+
+
+"""
+weight: 各种工作负载边划分权重
+按照点均衡划分
+"""
+
+
+def partition_data_for_tgnn(data: Data, num_parts: int, algo: str,
+                            verbose: bool = False,
+                            edge_weight_dict: dict = None):
+    if algo == "metis_for_tgnn":
+        part_fn = metis_for_tgnn
+
+    else:
+        raise ValueError(f"invalid algorithm: {algo}")
+    num_nodes = data.num_nodes
+    num_edges = data.num_edges
+    edge_index_dict = data.edge_index_dict
+    tgnn_norm = compute_temporal_norm(data.edge_index, data.edge_ts, num_nodes)
+    if verbose:
+        print(f"running partition algorithm: {algo}")
+    node_parts = part_fn(edge_index_dict, num_nodes, num_parts,
+                         edge_weight_dict)
+    edge_parts = node_parts[data.edge_index[1, :]]
+    eids = torch.arange(num_edges, dtype=torch.long)
+    data.eids = eids
+    data.sample_graph['eids'] = eids[data.sample_graph['eids']]
+    if data.y.dtype == torch.long:
+        if verbose:
+            print("compute num_classes")
+        num_classes = data.y.max().item() + 1
+    else:
+        num_classes = None
+
+    for i in range(num_parts):
+        npart_i = torch.where(node_parts == i)[0]        
+        epart_i = torch.where(edge_parts == i)[0]
+
+        pdata = {
+            "ids": npart_i,
+            "tgnn_norm": tgnn_norm,
+            "edge_index": data.edge_index[:, epart_i],
+            "sample_graph": data.sample_graph
+        } 
+        if num_classes is not None:
+            pdata["num_classes"] = num_classes
+    
+        for key, val in data:
+            if key == "edge_index" or key == "edge_index_dict" \
+                    or key == "sample_graph":
+                continue
+            if isinstance(val, torch.Tensor):
+                if val.size(0) == num_nodes:
+                    pdata[key] = val[npart_i]
+                elif val.size(0) == num_edges:
+                    pdata[key] = val[epart_i]
+                # else:
+                #     pdata[key] = val
+            elif isinstance(val, SparseTensor):
+                pass
+            else:
+                pdata[key] = val      
+        pdata = Data(**pdata)
+        yield pdata
+
+
+def metis_partition(edge_index, num_nodes: int, num_parts: int):
+    if num_parts <= 1:
+        return _nopart(edge_index, num_nodes)
+    G = nx.Graph()
+    G.add_nodes_from(torch.arange(0, num_nodes).tolist())
+    G.add_edges_from(edge_index.T.tolist())
+    nodes = torch.tensor(list(G.adj.keys()))
+    nodes = torch.tensor(list(G.adj.keys()))
+    cuts, part = metis.part_graph(G, num_parts)
+    node_parts = torch.zeros(num_nodes, dtype=torch.long)
+    node_parts[nodes] = torch.tensor(part)
+    edge_parts = node_parts[edge_index[1]]
+    return node_parts, edge_parts
+
+
+def metis_partition_bydegree(edge_index, num_nodes: int, num_parts: int):
+    if num_parts <= 1:
+        return _nopart(edge_index, num_nodes)
+    G = nx.Graph()
+    G.add_nodes_from(torch.arange(0, num_nodes).tolist())
+    G.add_edges_from(edge_index.T.tolist())
+    value, counts = torch.unique(edge_index[1, :].view(-1), return_counts=True)
+    nodes = torch.tensor(list(G.adj.keys()))
+    for i in range(value.shape[0]):
+        if (value[i].item() in G.nodes):
+            G.nodes[int(value[i].item())]['weight'] = counts[i]
+    G.graph['node_weight_attr'] = 'weight'
+    nodes = torch.tensor(list(G.adj.keys()))
+    cuts, part = metis.part_graph(G, num_parts)
+    node_parts = torch.zeros(num_nodes, dtype=torch.long)
+    node_parts[nodes] = torch.tensor(part)
+    edge_parts = node_parts[edge_index[1]]
+    return node_parts, edge_parts
+
+
+def compute_gcn_norm(edge_index: torch.LongTensor, num_nodes: int):
+    deg_j = degree(edge_index[0], num_nodes).pow(-0.5)
+    deg_i = degree(edge_index[1], num_nodes).pow(-0.5)
+    deg_i[deg_i.isinf() | deg_i.isnan()] = 0.0
+    deg_j[deg_j.isinf() | deg_j.isnan()] = 0.0
+    return deg_j[edge_index[0]] * deg_i[edge_index[1]]
+
+
+def compute_temporal_norm(edge_index: torch.LongTensor,
+                          timestamp: torch.FloatTensor,
+                          num_nodes: int):
+    srcavg, srcvar, dstavg, dstvar = get_norm_temporal(edge_index[0, :],
+                                                       edge_index[1, :],
+                                                       timestamp, num_nodes)
+    return srcavg, srcvar, dstavg, dstvar
+

starrygl/sample/part_utils/setup.py

+from setuptools import setup
+from torch.utils.cpp_extension import BuildExtension, CppExtension
+
+setup(
+    name='torch_utils',
+    ext_modules=[
+        CppExtension(
+            name='torch_utils',
+            sources=['./torch_utils.cpp'],
+            extra_compile_args=['-fopenmp','-Xlinker',' -export-dynamic'],
+            include_dirs=["../sample_core"],
+        ),
+    ],
+    cmdclass={
+        'build_ext': BuildExtension
+    })#
+
+#setup(
+#    name='cpu_cache_manager',
+#    ext_modules=[
+#        CppExtension(
+#            name='cpu_cache_manager',
+#            sources=['cpu_cache_manager.cpp'],
+#            extra_compile_args=['-fopenmp','-Xlinker',' -export-dynamic'],
+#            include_dirs=["./"],
+#        ),
+#    ],
+#    cmdclass={
+#        'build_ext': BuildExtension
+#    })#
+#
\ No newline at end of file

starrygl/sample/part_utils/torch_utils.cpp

+#include <pybind11/pybind11.h>
+#include <pybind11/numpy.h>
+#include <pybind11/stl.h>
+#include <torch/extension.h>
+#include <parallel_hashmap/phmap.h>
+#include <cstring>
+#include <vector>
+#include <iostream>
+#include <map>
+#include <boost/thread/mutex.hpp>
+using namespace std;
+#define MTX boost::mutex
+
+namespace py = pybind11;
+namespace th = torch;
+typedef int64_t NodeIDType;
+typedef float TimeStampType;
+
+#define EXTRAARGS , phmap::priv::hash_default_hash<K>, \
+                            phmap::priv::hash_default_eq<K>, \
+                            std::allocator<K>, 4, MTX
+template <class K>
+using HashT  = phmap::parallel_flat_hash_set<K EXTRAARGS>;
+
+template <class K, class V>
+using HashM  = phmap::parallel_flat_hash_map<K, V EXTRAARGS>;
+#define work_thread 10
+th::Tensor sparse_get_index(th::Tensor in,th::Tensor map_key){
+    auto key_ptr = map_key.data_ptr<NodeIDType>();
+    auto in_ptr = in.data_ptr<NodeIDType>();
+    int sz =  map_key.size(0);
+    vector<pair<NodeIDType,NodeIDType>> mp(sz);
+    vector<NodeIDType> out(in.size(0));
+#pragma omp parallel for
+    for(int i=0;i<sz;i++){
+        mp[i] = make_pair(key_ptr[i],i);
+    }
+    phmap::parallel_flat_hash_map<NodeIDType,NodeIDType> dict(mp.begin(),mp.end());
+
+#pragma omp parallel for
+    for(int i=0;i<in.size(0);i++){
+        out[i] = dict.find(in_ptr[i])->second;
+    }
+    return th::tensor(out);
+}
+vector<double> get_norm_temporal(th::Tensor row,th::Tensor col,th::Tensor timestamp,int num_nodes){
+    vector<double> ret(4);
+    HashM<NodeIDType,TimeStampType> dict0;
+    HashM<NodeIDType,TimeStampType> dict1;
+    auto rowptr = row.data_ptr<NodeIDType>();
+    auto colptr = col.data_ptr<NodeIDType>();
+    auto time_ptr = timestamp.data_ptr<TimeStampType>();
+    vector<TimeStampType> out_timestamp[work_thread];
+    vector<TimeStampType> in_timestamp[work_thread];
+#pragma omp parallel for num_threads(work_thread)
+    for(int i = 0;i<row.size(0);i++){
+        int tid = omp_get_thread_num();
+        if(dict0.find(rowptr[i])!=dict0.end()){
+            out_timestamp[tid].push_back(time_ptr[i]-dict0.find(rowptr[i])->second);
+            dict0.find(rowptr[i])->second = time_ptr[i];
+        }
+        else dict0.insert(make_pair(rowptr[i],time_ptr[i]));
+        if(dict1.find(colptr[i])!=dict1.end()){
+            in_timestamp[tid].push_back(time_ptr[i]-dict1.find(colptr[i])->second);
+            dict1.find(colptr[i])->second = time_ptr[i];
+        }
+        else dict1.insert(make_pair(colptr[i],time_ptr[i]));
+    }
+    double srcavg = 0;
+    double dstavg = 0;
+    double srcvar = 0;
+    double dstvar = 0;
+    double srccnt = 0;
+    double dstcnt = 0;
+    for(int i = 0;i<work_thread;i++){
+        #pragma omp parallel for num_threads(work_thread)
+        for(auto &v: in_timestamp[i]){
+            dstavg += v;
+            dstcnt++;
+        }
+        #pragma omp parallel for num_threads(work_thread)
+        for(auto &v: out_timestamp[i]){
+            srcavg += v;
+            srccnt++;
+        }
+    }
+    dstavg /= dstcnt;  
+    srcavg /= srccnt;
+    for(int i = 0;i<work_thread;i++){
+        #pragma omp parallel for num_threads(work_thread)
+        for(int j = 0;j<in_timestamp[i].size();j++){
+            TimeStampType v=in_timestamp[i][j];
+            dstvar += (v-dstavg)*(v-dstavg)/dstavg;
+        }
+        #pragma omp parallel for num_threads(work_thread)
+        for(int j = 0;j<out_timestamp[i].size();j++){
+            TimeStampType v=out_timestamp[i][j];
+            srcavg += (v-srcavg)*(v-srcavg)/srcavg;
+        }
+    }
+    ret[0]=srcavg;
+    ret[1]=srcvar;
+    ret[2]=dstavg;
+    ret[3]=dstvar;
+    return ret;
+}
+
+PYBIND11_MODULE(torch_utils, m)
+{
+    m
+    .def("sparse_get_index", 
+        &sparse_get_index, 
+        py::return_value_policy::reference)
+    .def("get_norm_temporal",
+        &get_norm_temporal, 
+        py::return_value_policy::reference
+    );
+}
\ No newline at end of file

starrygl/sample/sample_core/PreNegSampling.py

+import sys
+from os.path import abspath, join, dirname
+
+sys.path.insert(0, join(abspath(dirname(__file__))))
+from torch import Tensor
+import torch
+from base import NegativeSampling
+from base import NegativeSamplingMode
+from typing import Any, List, Optional, Tuple, Union
+
+    
+
+class PreNegativeSampling(NegativeSampling):
+    r"""The negative sampling configuration of a
+    :class:`~torch_geometric.sampler.BaseSampler` when calling
+    :meth:`~torch_geometric.sampler.BaseSampler.sample_from_edges`.
+    Args:
+        mode (str): The negative sampling mode
+            (:obj:`"binary"` or :obj:`"triplet"`).
+            If set to :obj:`"binary"`, will randomly sample negative links
+            from the graph.
+            If set to :obj:`"triplet"`, will randomly sample negative
+            destination nodes for each positive source node.
+        amount (int or float, optional): The ratio of sampled negative edges to
+            the number of positive edges. (default: :obj:`1`)
+        weight (torch.Tensor, optional): A node-level vector determining the
+            sampling of nodes. Does not necessariyl need to sum up to one.
+            If not given, negative nodes will be sampled uniformly.
+            (default: :obj:`None`)
+    """
+    def __init__(
+        self,
+        mode: Union[NegativeSamplingMode, str],
+        neg_sample_list: torch.Tensor
+    ):
+        super(PreNegativeSampling,self).__init__(mode)
+        self.neg_sample_list = neg_sample_list
+        self.next_pos = 0
+    def set_next_pos(self,pos):
+        self.next_pos = pos
+    def sample(self, num_samples: int,
+               num_nodes: Optional[int] = None) -> Tensor:
+        r"""Generates :obj:`num_samples` negative samples."""
+        if num_nodes is None:
+            raise ValueError(
+                f"Cannot sample negatives in '{self.__class__.__name__}' "
+                f"without passing the 'num_nodes' argument")
+        neg_sample_out = self.neg_sample_list[
+            self.next_pos:self.next_pos+num_samples,:].reshape(-1)
+        self.next_pos = self.next_pos + num_samples
+        return neg_sample_out
+        #return torch.from_numpy(np.random.randint(num_nodes, size=num_samples))

starrygl/sample/sample_core/Utils.py

+import os.path as osp
+import torch
+class GraphData():
+    def __init__(self, path):
+        assert path is not None and osp.exists(path),'path 不存在'
+        id,edge_index,data,partptr =torch.load(path)
+        # 当前分区序号
+        self.partition_id = id
+        # 总分区数
+        self.partitions = partptr.numel() - 1
+        # 全图结构数据
+        self.num_nodes = partptr[self.partitions]
+        self.num_edges = edge_index[0].numel()
+        self.edge_index = edge_index
+        # 该分区下的数据（包含特征向量和子图结构）pyg Data数据结构
+        self.data = data
+        # 分区映射关系
+        self.partptr = partptr
+        self.eid = [i for i in range(self.num_edges)]
+
+    def __init__(self, id, edge_index, data, partptr, timestamp=None):
+        # 当前分区序号
+        self.partition_id = id
+        # 总分区数
+        self.partitions = partptr.numel() - 1
+        # 全图结构数据
+        self.num_nodes = partptr[self.partitions]
+        if edge_index is not None:
+            self.num_edges = edge_index[0].numel()
+        self.edge_index = edge_index
+        self.edge_ts = timestamp
+        # 该分区下的数据（包含特征向量和子图结构）pyg Data数据结构
+        self.data = data
+        # 分区映射关系
+        self.partptr = partptr
+        # edge id
+        self.eid = torch.tensor([i for i in range(0, self.num_edges)])
+
+    def select_attr(self,index):
+        return torch.index_select(self.data.x,0,index)
+
+    #返回全局的节点id 所对应的分区
+    def get_part_num(self):
+        return self.data.x.size()[0]
+
+    def select_attr(self,index):
+        return torch.index_select(self.data.x,0,index)
+    def select_y(self,index):
+        return torch.index_select(self.data.y,0,index)
+    #返回全局的节点id 所对应的分区
+    def get_localId_by_partitionId(self,id,index):
+        #print(index)
+        if(id == -1 or id == 0):
+            return index
+        else:
+            return torch.add(index,-self.partptr[id])
+    def get_globalId_by_partitionId(self,id,index):
+        if(id == -1 or id == 0):
+            return index
+        else:
+            return torch.add(index,self.partptr[id])
+
+    def get_node_num(self):    
+        return self.num_nodes
+    
+    
+    def localId_to_globalId(self,id,partitionId:int = -1):
+        '''
+        将分区partitionId内的点id映射为全局的id
+        '''
+        if partitionId == -1:
+            partitionId = self.partition_id
+        assert id >=self.partptr[self.partition_id] and id < self.partptr[self.partition_id+1]
+        ids_before = 0
+        if self.partition_id>0:
+            ids_before = self.partptr[self.partition_id-1]
+        return id+ids_before
+    
+    def get_partitionId_by_globalId(self,id):
+        '''
+        通过全局id得到对应的分区序号
+        '''
+        partitionId = -1
+        assert id>=0 and id<self.num_nodes,'id 超过范围'
+        for i in range(self.partitions):
+            if id>=self.partptr[i] and id<self.partptr[i+1]:
+                partitionId = i
+                break
+        assert partitionId>=0, 'id 不存在对应的分区'
+        return partitionId
+    
+    def get_nodes_by_partitionId(self,id):
+        '''
+        根据partitioId 返回该分区的节点数量
+        
+        '''
+        assert id>=0 and id<self.partitions,'partitionId 非法'
+        return (int)(self.partptr[id+1]-self.partptr[id])
+        
+    
+    def __repr__(self):
+        return (f'{self.__class__.__name__}(\n'
+                f'  partition_id={self.partition_id}\n'
+                f'  data={self.data},\n'
+                f'  global_info('
+                f'num_nodes={self.num_nodes},'
+                f' num_edges={self.num_edges},'
+                f' num_parts={self.partitions},'
+                f' edge_index=[2,{self.edge_index[0].numel()}])\n'
+                f')')

starrygl/sample/sample_core/base.py

+import torch
+from torch import Tensor
+from enum import Enum
+import math
+from abc import ABC
+from typing import Any, List, Optional, Tuple, Union
+
+class SampleType(Enum):
+    Whole = 0
+    Inner = 1
+    Outer =2
+
+
+class NegativeSamplingMode(Enum):
+    # 'binary': Randomly sample negative edges in the graph.
+    binary = 'binary'
+    # 'triplet': Randomly sample negative destination nodes for each positive
+    # source node.
+    triplet = 'triplet'
+    
+
+class NegativeSampling:
+    r"""The negative sampling configuration of a
+    :class:`~torch_geometric.sampler.BaseSampler` when calling
+    :meth:`~torch_geometric.sampler.BaseSampler.sample_from_edges`.
+    Args:
+        mode (str): The negative sampling mode
+            (:obj:`"binary"` or :obj:`"triplet"`).
+            If set to :obj:`"binary"`, will randomly sample negative links
+            from the graph.
+            If set to :obj:`"triplet"`, will randomly sample negative
+            destination nodes for each positive source node.
+        amount (int or float, optional): The ratio of sampled negative edges to
+            the number of positive edges. (default: :obj:`1`)
+        weight (torch.Tensor, optional): A node-level vector determining the
+            sampling of nodes. Does not necessariyl need to sum up to one.
+            If not given, negative nodes will be sampled uniformly.
+            (default: :obj:`None`)
+    """
+    mode: NegativeSamplingMode
+    amount: Union[int, float] = 1
+    weight: Optional[Tensor] = None
+    unique: bool
+
+    def __init__(
+        self,
+        mode: Union[NegativeSamplingMode, str],
+        amount: Union[int, float] = 1,
+        weight: Optional[Tensor] = None,
+        unique: bool = False
+    ):
+        self.mode = NegativeSamplingMode(mode)
+        self.amount = amount
+        self.weight = weight
+        self.unique = unique
+
+        if self.amount <= 0:
+            raise ValueError(f"The attribute 'amount' needs to be positive "
+                             f"for '{self.__class__.__name__}' "
+                             f"(got {self.amount})")
+
+        if self.is_triplet():
+            if self.amount != math.ceil(self.amount):
+                raise ValueError(f"The attribute 'amount' needs to be an "
+                                 f"integer for '{self.__class__.__name__}' "
+                                 f"with 'triplet' negative sampling "
+                                 f"(got {self.amount}).")
+            self.amount = math.ceil(self.amount)
+
+    def is_binary(self) -> bool:
+        return self.mode == NegativeSamplingMode.binary
+
+    def is_triplet(self) -> bool:
+        return self.mode == NegativeSamplingMode.triplet
+
+    def sample(self, num_samples: int,
+               num_nodes: Optional[int] = None) -> Tensor:
+        r"""Generates :obj:`num_samples` negative samples."""
+        if self.weight is None:
+            if num_nodes is None:
+                raise ValueError(
+                    f"Cannot sample negatives in '{self.__class__.__name__}' "
+                    f"without passing the 'num_nodes' argument")
+            return torch.randint(num_nodes, (num_samples, ))
+
+        if num_nodes is not None and self.weight.numel() != num_nodes:
+            raise ValueError(
+                f"The 'weight' attribute in '{self.__class__.__name__}' "
+                f"needs to match the number of nodes {num_nodes} "
+                f"(got {self.weight.numel()})")
+        return torch.multinomial(self.weight, num_samples, replacement=True)
+
+class SampleOutput:
+    node: Optional[torch.Tensor] = None
+    edge_index_list: Optional[List[torch.Tensor]] = None
+    eid_list: Optional[List[torch.Tensor]] = None
+    delta_ts_list: Optional[List[torch.Tensor]] = None
+    metadata: Optional[Any] = None
+
+class BaseSampler(ABC):
+    r"""An abstract base class that initializes a graph sampler and provides
+    :meth:`_sample_one_layer_from_nodes`
+    :meth:`_sample_one_layer_from_nodes_parallel`
+    :meth:`sample_from_nodes` routines.
+    """
+
+    def sample_from_nodes(
+        self,
+        nodes: torch.Tensor,
+        with_outer_sample: SampleType,
+        **kwargs
+    ) -> Tuple[torch.Tensor, list]:
+        r"""Performs mutilayer sampling from the nodes specified in: nodes
+        The specific number of layers is determined by parameter: num_layers
+        returning a sampled subgraph in the specified output format: Tuple[torch.Tensor, list].
+
+        Args:
+            nodes: the list of seed nodes index
+            with_outer_sample: 0-sample in whole graph structure; 1-sample onehop outer nodel; 2-cross partition sampling
+            **kwargs: other kwargs
+        Returns:
+            sampled_nodes: the nodes sampled
+            sampled_edge_index_list: the edges sampled
+        """
+        raise NotImplementedError
+    
+    def sample_from_edges(
+        self,
+        edges: torch.Tensor,
+        with_outer_sample: SampleType,
+        edge_label: Optional[torch.Tensor] = None,
+        neg_sampling: Optional[NegativeSampling] = None
+    ) -> Tuple[torch.Tensor, list]:
+        r"""Performs sampling from the edges specified in :obj:`index`,
+        returning a sampled subgraph in the specified output format.
+
+        Args:
+            edges: the list of seed edges index
+            with_outer_sample: 0-sample in whole graph structure; 1-sample onehop outer nodel; 2-cross partition sampling
+            edge_label: the label for the seed edges.
+            neg_sampling: The negative sampling configuration
+        Returns:
+            sampled_nodes: the nodes sampled
+            sampled_edge_index_list: the edges sampled
+            metadata: other infomation
+        """
+        raise NotImplementedError
+
+    # def _sample_one_layer_from_nodes(
+    #     self,
+    #     nodes:torch.Tensor,
+    #     **kwargs
+    # ) -> Tuple[torch.Tensor, torch.Tensor]:
+    #     r"""Performs sampling from the nodes specified in: nodes,
+    #     returning a sampled subgraph in the specified output format: Tuple[torch.Tensor, torch.Tensor].
+
+    #     Args:
+    #         nodes: the list of seed nodes index
+    #         **kwargs: other kwargs
+    #     Returns:
+    #         sampled_nodes: the nodes sampled
+    #         sampled_edge_index: the edges sampled
+    #     """
+    #     raise NotImplementedError
+    
+    # def _sample_one_layer_from_nodes_parallel(
+    #     self, 
+    #     nodes: torch.Tensor,
+    #     **kwargs
+    # ) -> Tuple[torch.Tensor, torch.Tensor]:
+    #     r"""Performs sampling paralleled from the nodes specified in: nodes,
+    #     returning a sampled subgraph in the specified output format: Tuple[torch.Tensor, torch.Tensor].
+
+    #     Args:
+    #         nodes: the list of seed nodes index
+    #         **kwargs: other kwargs
+    #     Returns:
+    #         sampled_nodes: the nodes sampled
+    #         sampled_edge_index: the edges sampled
+    #     """
+    #     raise NotImplementedError

starrygl/sample/sample_core/distparser.py

+
+from enum import Enum
+import sys
+import argparse
+from os.path import abspath, join, dirname
+import time
+
+sys.path.insert(0, join(abspath(dirname(__file__))))
+class SampleType(Enum):
+    Whole = 0
+    Inner = 1
+    Outer =2
+
+parser = argparse.ArgumentParser(
+    description="RPC Reinforcement Learning Example",
+    formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+)
+parser.add_argument('--world_size', default=1, type=int, metavar='W',
+                    help='number of workers')
+parser.add_argument('--rank', default=0, type=int, metavar='W',
+                    help='rank of the worker')
+parser.add_argument('--log_interval', type=int, default=10, metavar='N',
+                    help='interval between training status logs')
+parser.add_argument('--gamma', type=float, default=0.99, metavar='G',
+                    help='how much to value future rewards')
+parser.add_argument('--seed', type=int, default=1, metavar='S',
+                    help='random seed  for reproducibility')
+parser.add_argument('--num_sampler', type=int, default=1, metavar='S',
+                    help='number of samplers')
+parser.add_argument('--queue_size', type=int, default=10000, metavar='S',
+                    help='sampler queue size')
+#parser = distparser.parser.add_subparsers().add_parser("train")#argparse.ArgumentParser(description='minibatch_gnn_models')
+parser.add_argument('--data', type=str, help='dataset name')
+parser.add_argument('--config', type=str, help='path to config file')
+parser.add_argument('--gpu', type=str, default='0', help='which GPU to use')
+parser.add_argument('--model_name', type=str, default='', help='name of stored model')
+parser.add_argument('--rand_edge_features', type=int, default=0, help='use random edge featrues')
+parser.add_argument('--rand_node_features', type=int, default=0, help='use random node featrues')
+parser.add_argument('--eval_neg_samples', type=int, default=1, help='how many negative samples to use at inference. Note: this will change the metric of test set to AP+AUC to AP+MRR!')
+args = parser.parse_args()
+rpc_proxy=None
+WORKER_RANK = args.rank
+NUM_SAMPLER = args.num_sampler
+WORLD_SIZE = args.world_size
+QUEUE_SIZE = args.queue_size
+MAX_QUEUE_SIZE = 5*args.queue_size
+RPC_NAME = "rpcserver{}"
+SAMPLE_TYPE = SampleType.Outer
+def _get_worker_rank():
+    return WORKER_RANK
+def _get_num_sampler():
+    return NUM_SAMPLER
+def _get_world_size():
+    return WORLD_SIZE
+def _get_RPC_NAME():
+    return RPC_NAME
+def _get_queue_size():
+    return QUEUE_SIZE
+def _get_max_queue_size():
+    return MAX_QUEUE_SIZE
+def _get_rpc_name():
+    return RPC_NAME
\ No newline at end of file

starrygl/sample/sample_core/graph_store.py

+import sys
+from os.path import abspath, join, dirname
+import time
+
+sys.path.insert(0, join(abspath(dirname(__file__))))
+graph_set={}
+
+def _clear_all(barrier = None):
+    global graph_set
+    for key in graph_set:
+        graph = graph_set[key]
+        graph._close_graph_in_shame()
+        print('clear ',key)
+        if(barrier is not None and barrier.wait()==0):
+            graph._unlink_graph_in_shame()
+    graph_set = {}
+
+def _set_graph(graph_name,graph_info):
+    graph_info._get_graph_from_shm()
+    graph_set[graph_name]=graph_info
+
+def _get_graph(graph_name):
+    return graph_set[graph_name]
+
+def _del_graph(graph_name):
+    graph_set.pop(graph_name)
+
+def _get_size():
+    return len(graph_set)
+
+
+# local_sampler=None
+local_sampler = {}
+def set_local_sampler(graph_name,sampler):
+    local_sampler[graph_name] = sampler
+
+def get_local_sampler(sampler_name):
+    
+    assert sampler_name in local_sampler, 'Local_sampler doesn\'t has sampler_name'
+    return local_sampler[sampler_name]
\ No newline at end of file

starrygl/sample/sample_core/my_sample_test_before.py

+import torch
+import time
+from Utils import GraphData
+seed = 10  # 你可以选择任何整数作为种子
+torch.manual_seed(seed)
+
+num_nodes1 = 10
+fanout1 = [2]
+edge_index1 = torch.tensor([[1, 5, 7, 9, 2, 4, 6, 7, 8, 0, 1, 6, 2, 0, 1, 3, 5, 8, 9, 7, 4, 8, 2, 3, 5, 8],
+                            [0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 3, 3, 4, 4, 4, 4, 4, 5, 6, 6, 7, 7, 8, 9]])
+edge_ts =      torch.tensor([1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 4, 4, 4, 4, 4, 5, 5, 5, 6, 6, 6, 6]).double()
+edge_weight1 = torch.tensor([2, 1, 2, 1, 8, 6, 3, 1, 1, 1, 1, 5, 1, 1, 2, 1, 1, 1, 1, 5, 1, 2, 2, 2, 1, 1]).double()
+edge_weight1 = None
+
+g_data = GraphData(id=0, edge_index=edge_index1, timestamp=edge_ts, data=None, partptr=torch.tensor([0, num_nodes1]))
+from neighbor_sampler import NeighborSampler, SampleType
+pre = time.time()
+# from neighbor_sampler import get_neighbors, update_edge_weight
+# row, col = edge_index1
+# tnb = get_neighbors(row.contiguous(), col.contiguous(), num_nodes1, edge_weight1)
+# print("tnb.neighbors:", tnb.neighbors)
+# print("tnb.deg:", tnb.deg)
+# print("tnb.weight:", tnb.edge_weight)
+sampler = NeighborSampler(num_nodes1, 
+                          num_layers=1, 
+                          fanout=fanout1, 
+                          edge_weight=edge_weight1, 
+                          graph_data=g_data, 
+                          workers=2, 
+                          graph_name='a',
+                          is_distinct = 0,
+                          policy="recent")
+end = time.time()
+print("init time:", end-pre)
+print("tnb.neighbors:", sampler.tnb.neighbors)
+print("tnb.deg:", sampler.tnb.deg)
+print("tnb.ts:", sampler.tnb.timestamp)
+print("tnb.weight:", sampler.tnb.edge_weight)
+
+
+# update_edge_row =     row
+# update_edge_col =     col
+# update_edge_w =  torch.DoubleTensor([i for i in range(edge_weight1.size(0))])
+# print('tnb.edge_weight:', tnb.edge_weight)
+# print('begin update')
+# pre = time.time()
+# update_edge_weight(tnb, update_edge_row.contiguous(), update_edge_col.contiguous(), update_edge_w.contiguous())
+# end = time.time()
+# print("update time:", end-pre)
+# print('update_edge_row:', update_edge_row)
+# print('update_edge_col:', update_edge_col)
+# print('tnb.edge_weight:', tnb.edge_weight)
+
+
+pre = time.time()
+
+out = sampler.sample_from_nodes(torch.tensor([6,7]),
+                                 with_outer_sample=SampleType.Whole, 
+                                 ts=torch.tensor([9, 9]))
+end = time.time()
+# print('node:', out.node)
+# print('edge_index_list:', out.edge_index_list)
+# print('eid_list:', out.eid_list)
+# print('eid_ts_list:', out.eid_ts_list)
+print("sample time:", end-pre)
+print("tot_time", out[0].tot_time)
+print("sam_time", out[0].sample_time)
+print("sam_edge", out[0].sample_edge_num)
+print('eid_list:', out[0].eid)
+print('delta_ts_list:', out[0].delta_ts)
+print((out[0].sample_nodes<10000).sum())
+print('node:', out[0].sample_nodes)
+print('node_ts:', out[0].sample_nodes_ts)
\ No newline at end of file

starrygl/sample/sample_core/my_sample_test_static.py

+import torch
+import time
+from Utils import GraphData
+seed = 10  # 你可以选择任何整数作为种子
+torch.manual_seed(seed)
+
+num_nodes1 = 10
+fanout1 = [2,2]      # index 0  1  2  3  4  5  6  7  8  9  10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
+edge_index1 = torch.tensor([[1, 5, 7, 9, 2, 4, 6, 7, 8, 0, 1, 6, 2, 0, 1, 3, 5, 8, 9, 7, 4, 8, 2, 3, 5, 8],
+                            [0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 3, 3, 4, 4, 4, 4, 4, 5, 6, 6, 7, 7, 8, 9]])
+edge_weight1 = torch.tensor([2, 1, 2, 1, 8, 6, 3, 1, 1, 1, 1, 5, 1, 1, 2, 1, 1, 1, 1, 5, 1, 2, 2, 2, 1, 1]).double()
+
+g_data = GraphData(id=0, edge_index=edge_index1, data=None, partptr=torch.tensor([0, num_nodes1]))
+
+edge_weight1 = None
+# g_data.eid=None
+
+from neighbor_sampler import NeighborSampler, SampleType
+pre = time.time()
+sampler = NeighborSampler(num_nodes1, 
+                          num_layers=2, 
+                          fanout=fanout1, 
+                          edge_weight=edge_weight1,
+                          graph_data=g_data, 
+                          workers=2, 
+                          graph_name='a',
+                          policy="uniform")
+end = time.time()
+print("init time:", end-pre)
+
+print("tnb.neighbors:", sampler.tnb.neighbors)
+print("tnb.eid:", sampler.tnb.eid)
+print("tnb.deg:", sampler.tnb.deg)
+print("tnb.weight:", sampler.tnb.edge_weight)
+
+# row,col = edge_index1
+# update_edge_row =     row
+# update_edge_col =     col
+# update_edge_w =  torch.FloatTensor([i for i in range(edge_weight1.size(0))])
+# print('tnb.edge_weight:', sampler.tnb.edge_weight)
+# print('begin update')
+# pre = time.time()
+# sampler.tnb.update_edge_weight(sampler.tnb, update_edge_row.contiguous(), update_edge_col.contiguous(), update_edge_w.contiguous())
+# end = time.time()
+# print("update time:", end-pre)
+# print('update_edge_row:', update_edge_row)
+# print('update_edge_col:', update_edge_col)
+# print('tnb.edge_weight:', sampler.tnb.edge_weight)
+
+
+pre = time.time()
+
+out = sampler.sample_from_nodes(torch.tensor([1,2]), 
+                                with_outer_sample=SampleType.Whole)# sampler.sample_from_nodes(torch.masked_select(torch.arange(g.num_nodes),node_data['train_mask']))
+
+end = time.time()
+print('node1:\t', out[0].sample_nodes().tolist())
+print('eid1:\t', out[0].eid().tolist())
+print('edge1:\t', edge_index1[:, out[0].eid()].tolist())
+print('node2:\t', out[1].sample_nodes().tolist())
+print('eid2:\t', out[1].eid().tolist())
+print('edge2:\t', edge_index1[:, out[1].eid()].tolist())
+print("sample time:", end-pre)
\ No newline at end of file

starrygl/sample/sample_core/neighbor_sampler.py

+import sys
+from os.path import abspath, join, dirname
+
+sys.path.insert(0, join(abspath(dirname(__file__))))
+import math
+import torch
+import torch.multiprocessing as mp
+from typing import Optional, Tuple
+
+import graph_store
+from distparser import SampleType, NUM_SAMPLER
+from base import BaseSampler, NegativeSampling, SampleOutput
+from sample_cores import ParallelSampler, get_neighbors, heads_unique
+from torch.distributed.rpc import rpc_async
+
+def outer_sample(graph_name, nodes, ts, fanout_index, with_outer_sample = SampleType.Outer):# 默认此时继续向外采样
+    local_sampler = graph_store.get_local_sampler(graph_name)
+    assert local_sampler is not None, 'Local_sampler is None!!!'
+    out = local_sampler.sample_from_nodes(nodes, with_outer_sample, ts, fanout_index)
+    return out
+
+class NeighborSampler(BaseSampler):
+    def __init__(
+        self,
+        num_nodes: int,
+        num_layers: int,
+        fanout: list,
+        graph_data,
+        workers = 1,
+        tnb = None,
+        is_distinct = 0,
+        policy = "uniform",
+        edge_weight: Optional[torch.Tensor] = None,
+        graph_name = None
+    ) -> None:
+        r"""__init__
+        Args:
+            num_nodes: the num of all nodes in the graph
+            num_layers: the num of layers to be sampled
+            fanout: the list of max neighbors' number chosen for each layer
+            workers: the number of threads, default value is 1
+            tnb: neighbor infomation table
+            is_distinct: 1-need distinct muti-edge, 0-don't need distinct muti-edge
+            policy: "uniform" or "recent" or "weighted"
+            edge_weight: the initial weights of edges
+            graph_name: the name of graph
+        should provide edge_index or (neighbors, deg)
+        """
+        super().__init__()
+        self.num_layers = num_layers
+        # 线程数不超过torch默认的omp线程数
+        self.workers = workers # min(workers, torch.get_num_threads())
+        self.fanout = fanout
+        self.num_nodes = num_nodes
+        self.graph_data=graph_data
+        self.policy = policy
+        self.is_distinct = is_distinct
+        assert graph_name is not None
+        self.graph_name = graph_name
+        if(tnb is None):
+            if(graph_data.edge_ts is not None):
+                timestamp,ind = graph_data.edge_ts.sort()
+                timestamp = timestamp.float().contiguous()
+                eid = graph_data.eid[ind].contiguous()
+                row, col = graph_data.edge_index[:,ind]
+            else:
+                eid = graph_data.eid
+                timestamp = None
+                row, col = graph_data.edge_index
+            if(edge_weight is not None):
+                edge_weight = edge_weight.float().contiguous()
+            self.tnb = get_neighbors(graph_name, row.contiguous(), col.contiguous(), num_nodes, is_distinct, eid, edge_weight, timestamp)
+        else:
+            assert tnb is not None
+            self.tnb = tnb
+        self.p_sampler = ParallelSampler(self.tnb, num_nodes, graph_data.num_edges, workers, 
+                                         fanout, num_layers, policy)
+    
+    def _get_sample_info(self):
+        return self.num_nodes,self.num_layers,self.fanout,self.workers
+    
+    def _get_sample_options(self):
+        return {"is_distinct" : self.is_distinct,
+                "policy" : self.policy,
+                "with_eid" : self.tnb.with_eid, 
+                "weighted" : self.tnb.weighted, 
+                "with_timestamp" : self.tnb.with_timestamp}
+    
+    def insert_edges_with_timestamp(
+            self, 
+            edge_index : torch.Tensor, 
+            eid : torch.Tensor, 
+            timestamp : torch.Tensor,
+            edge_weight : Optional[torch.Tensor] = None):
+        row, col = edge_index
+        # 更新节点数和tnb
+        self.num_nodes = self.tnb.update_neighbors_with_time(
+            row.contiguous(), 
+            col.contiguous(), 
+            timestamp.contiguous(), 
+            eid.contiguous(), 
+            self.is_distinct, 
+            edge_weight.contiguous())
+    
+    def update_edges_weight(
+            self, 
+            edge_index : torch.Tensor, 
+            eid : torch.Tensor,
+            edge_weight : Optional[torch.Tensor] = None):
+        row, col = edge_index
+        # 更新tnb的权重信息
+        if self.tnb.with_eid:
+            self.tnb.update_edge_weight(
+                eid.contiguous(),
+                col.contiguous(),
+                edge_weight.contiguous()
+            )
+        else:
+            self.tnb.update_edge_weight(
+                row.contiguous(),
+                col.contiguous(),
+                edge_weight.contiguous()
+            )
+    
+    def update_nodes_weight(
+            self, 
+            nid : torch.Tensor,
+            node_weight : Optional[torch.Tensor] = None):
+        # 更新tnb的权重信息
+        self.tnb.update_node_weight(
+            nid.contiguous(),
+            node_weight.contiguous()
+        )
+
+    def update_all_node_weight(
+            self,
+            node_weight : torch.Tensor):
+        # 更新tnb的权重信息
+        self.tnb.update_all_node_weight(node_weight.contiguous())
+
+    def sample_from_nodes(
+        self,
+        nodes: torch.Tensor,
+        ts: Optional[torch.Tensor] = None,
+        with_outer_sample: SampleType = SampleType.Whole
+    ) -> SampleOutput:
+        r"""Performs mutilayer sampling from the nodes specified in: nodes
+        The specific number of layers is determined by parameter: num_layers
+        returning a sampled subgraph in the specified output format: Tuple[torch.Tensor, list].
+
+        Args:
+            nodes: the list of seed nodes index,
+            ts: the timestamp of nodes, optional,
+            with_outer_sample: 0-sample in whole graph structure; 1-sample onehop outer nodel; 2-cross partition sampling
+            fanout_index: optional. Specify the index to fanout
+        Returns:
+            sampled_nodes: the node sampled
+            sampled_edge_index_list: the edge sampled
+        """
+        if(ts is None):
+            self.p_sampler.neighbor_sample_from_nodes(nodes.contiguous(), None)
+            ret = self.p_sampler.get_ret()
+            return ret
+        else:
+            self.p_sampler.neighbor_sample_from_nodes(nodes.contiguous(), ts.float().contiguous())
+            ret = self.p_sampler.get_ret()
+            return ret
+        
+    def sample_from_edges(
+        self,
+        edges: torch.Tensor,
+        ets: Optional[torch.Tensor] = None,
+        neg_sampling: Optional[NegativeSampling] = None,
+        with_outer_sample: SampleType = SampleType.Whole
+    ) -> SampleOutput:
+        r"""Performs sampling from the edges specified in :obj:`index`,
+        returning a sampled subgraph in the specified output format.
+
+        Args:
+            edges: the list of seed edges index
+            with_outer_sample: 0-sample in whole graph structure; 1-sample onehop outer nodel; 2-cross partition sampling
+            ets: the timestamp of edges, optional
+            neg_sampling: The negative sampling configuration
+        Returns:
+            sampled_edge_index_list: the edges sampled
+            sampled_eid_list: the edges' id sampled
+            sampled_delta_ts_list:the edges' delta time sampled
+            metadata: other infomation
+        """
+        src, dst = edges
+        num_pos = src.numel()
+        num_neg = 0
+        with_timestap = ets is not None
+        seed_ts = None
+
+        if neg_sampling is not None:
+            num_neg = math.ceil(num_pos * neg_sampling.amount)
+            if neg_sampling.is_binary():
+                src_neg = neg_sampling.sample(num_neg, self.num_nodes)
+                dst_neg = neg_sampling.sample(num_neg, self.num_nodes)
+                seed = torch.cat([src, dst, src_neg, dst_neg], dim=0)
+                if with_timestap: # ts操作
+                    seed_ts = torch.cat([ets, ets, ets, ets], dim=0)
+            if neg_sampling.is_triplet():
+                src_neg = neg_sampling.sample(num_neg, self.num_nodes)
+                seed = torch.cat([src, dst, src_neg], dim=0)
+                if with_timestap: # ts操作
+                    seed_ts = torch.cat([ets, ets, ets], dim=0)
+        else:
+            seed = torch.cat([src, dst], dim=0)            
+            if with_timestap: # ts操作
+                seed_ts = torch.cat([ets, ets], dim=0)
+
+        # 去重负采样
+        if neg_sampling is not None and neg_sampling.unique:
+            if with_timestap: # ts操作
+                pair, inverse_seed= torch.unique(torch.stack([seed, seed_ts],0), return_inverse=True, dim=1)
+                seed, seed_ts = pair
+                seed = seed.long()
+            else:
+                seed, inverse_seed = seed.unique(return_inverse=True)
+
+        out = self.sample_from_nodes(seed, seed_ts, with_outer_sample)
+        
+        if neg_sampling is None or (not neg_sampling.unique):
+            if with_timestap:
+                return out, {'seed':seed,'seed_ts':seed_ts,
+                             'src_pos_index':torch.arange(0,num_pos), 
+                             'dst_pos_index':torch.arange(num_pos,2*num_pos), 
+                             'src_neg_index':torch.arange(2*num_pos,3*num_pos)}
+            else:
+                return out, {'seed':seed,
+                             'src_pos_index':slice(0,num_pos), 
+                             'dst_pos_index':slice(num_pos,2*num_pos), 
+                             'src_neg_index':slice(2*num_pos,3*num_pos)}
+
+        metadata = {}
+        if neg_sampling.is_binary():
+            src_pos_index = inverse_seed[:num_pos]
+            dst_pos_index = inverse_seed[num_pos:2 * num_pos]
+            src_neg_index = inverse_seed[2 * num_pos:3 * num_pos]
+            src_neg_index = src_neg_index.view(num_pos, -1).squeeze(-1)
+            dst_neg_index = inverse_seed[3 * num_pos:]
+            dst_neg_index = dst_neg_index.view(num_pos, -1).squeeze(-1)
+            metadata = {'seed':seed, 'src_neg_index':src_neg_index, 'dst_pos_index':dst_pos_index, 'dst_neg_index':dst_neg_index}
+            if with_timestap:
+                metadata['seed_ts'] = seed_ts
+        elif neg_sampling.is_triplet():
+            src_pos_index = inverse_seed[:num_pos]
+            dst_pos_index = inverse_seed[num_pos:2 * num_pos]
+            src_neg_index = inverse_seed[2 * num_pos:]
+            src_neg_index = src_neg_index.view(num_pos, -1).squeeze(-1)
+            # src_index是seed里src点的索引
+            # dst_pos_index是seed里dst_pos点的索引
+            # dst_neg_index是seed里dst_neg点的索引
+            metadata = {'seed':seed, 'src_pos_index':src_pos_index, 'src_neg_index':src_neg_index, 'dst_pos_index':dst_pos_index}
+            if with_timestap:
+                metadata['seed_ts'] = seed_ts
+        # sampled_nodes最前方是原始序列的采样起点也就是去重后的seed
+        return out, metadata
+    
+
+
+if __name__=="__main__":
+    # edge_index1 = torch.tensor([[0, 1, 1, 1, 2, 2, 2, 4, 4, 4, 5], # , 3, 3
+    #                             [1, 0, 2, 4, 1, 3, 0, 3, 5, 0, 2]])# , 2, 5
+    edge_index1 = torch.tensor([[0, 1, 1, 1, 1, 2, 2, 2, 2, 4, 4, 4, 5], # , 3, 3
+                                [1, 0, 2, 0, 4, 1, 3, 0, 3, 3, 5, 0, 2]])# , 2, 5
+    edge_weight1 = None
+    timeStamp=torch.FloatTensor([1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4])
+    num_nodes1 = 6
+    num_neighbors = 2
+    # Run the neighbor sampling
+    from Utils import GraphData
+    g_data = GraphData(id=0, edge_index=edge_index1, timestamp=timeStamp, data=None, partptr=torch.tensor([0, num_nodes1]))
+    sampler = NeighborSampler(num_nodes=num_nodes1, 
+                              num_layers=3, 
+                              fanout=[2, 1, 1], 
+                              edge_weight=edge_weight1, 
+                              graph_data=g_data, 
+                              graph_name='a',
+                              workers=4,
+                              is_distinct = 0)
+
+    out = sampler.sample_from_nodes(torch.tensor([1,2]),
+                                    ts=torch.tensor([1, 2]),
+                                    with_outer_sample=SampleType.Whole)
+    # out = sampler.sample_from_edges(torch.tensor([[1,2],[4,0]]), 
+    #                                 with_outer_sample=SampleType.Whole, 
+    #                                 ets = torch.tensor([1, 2]))
+    
+    # Print the result
+    print('node:', out.node)
+    print('edge_index_list:', out.edge_index_list)
+    print('eid_list:', out.eid_list)
+    print('delta_ts_list:', out.delta_ts_list)
+    print('metadata: ', out.metadata)

starrygl/sample/sample_core/ogb_demo_before.py

+import torch
+from ogb.nodeproppred import PygNodePropPredDataset
+from torch_geometric import datasets
+import time
+from Utils import GraphData
+
+def load_ogb_dataset(name, data_path):
+    dataset = PygNodePropPredDataset(name=name, root=data_path)
+    split_idx = dataset.get_idx_split()
+    g = dataset[0]
+    n_node = g.num_nodes
+    node_data={}
+    node_data['train_mask'] = torch.zeros(n_node, dtype=torch.bool)
+    node_data['val_mask'] = torch.zeros(n_node, dtype=torch.bool)
+    node_data['test_mask'] = torch.zeros(n_node, dtype=torch.bool)
+    node_data['train_mask'][split_idx["train"]] = True
+    node_data['val_mask'][split_idx["valid"]] = True
+    node_data['test_mask'][split_idx["test"]] = True
+    return g, node_data
+
+g, node_data = load_ogb_dataset('ogbn-products', "/home/zlj/hzq/code/gnn/dataset/")
+print(g)
+# for worker in [1,2,3,4,5,6,7,8,9,10,20,30]:
+# import random
+# timestamp = [random.randint(1, 5) for i in range(0, g.num_edges)]
+# timestamp = torch.FloatTensor(timestamp)
+
+print('begin load')
+pre = time.time()
+timestamp = torch.load('/home/zlj/hzq/code/gnn/my_sampler/TemporalSample/timestamp.my')
+tnb = torch.load("tnb_before.my")
+end = time.time()
+print("load time:", end-pre)
+row, col = g.edge_index
+edge_weight=None
+g_data = GraphData(id=1, edge_index=g.edge_index, timestamp=timestamp, data=g, partptr=torch.tensor([0, g.num_nodes//4, g.num_nodes//4*2, g.num_nodes//4*3, g.num_nodes]))
+
+from neighbor_sampler import NeighborSampler, SampleType, get_neighbors
+
+# print('begin tnb')
+# pre = time.time()
+# tnb = get_neighbors(row.contiguous(), col.contiguous(), g.num_nodes, 0, g_data.eid, edge_weight, timestamp)
+# end = time.time()
+# print("init tnb time:", end-pre)
+# torch.save(tnb, "tnb_before.my")
+
+
+pre = time.time()
+sampler = NeighborSampler(g.num_nodes, 
+                          tnb=tnb,
+                          num_layers=2, 
+                          fanout=[100,100], 
+                          graph_data=g_data, 
+                          workers=10, 
+                          policy="uniform", 
+                          is_root_ts=0, 
+                          graph_name='a')
+end = time.time()
+print("init time:", end-pre)
+
+# from torch_geometric.sampler import NeighborSampler, NumNeighbors, NodeSamplerInput, SamplerOutput
+# pre = time.time()
+# num_nei = NumNeighbors([100, 100])
+# node_idx = NodeSamplerInput(input_id=None, node=torch.tensor(range(g.num_nodes//4, g.num_nodes//4+600000)))# (input_id=None, node=torch.masked_select(torch.arange(g.num_nodes),node_data['train_mask']))
+# sampler = NeighborSampler(g, num_nei)
+# end = time.time()
+# print("init time:", end-pre)
+
+ts = torch.tensor([i%5+1 for i in range(0, 600000)])
+pre = time.time()
+out = sampler.sample_from_nodes(torch.tensor(range(g.num_nodes//4, g.num_nodes//4+600000)), 
+                                ts=ts, 
+                                with_outer_sample=SampleType.Inner)# sampler.sample_from_nodes(torch.masked_select(torch.arange(g.num_nodes),node_data['train_mask']))
+
+# out = sampler.sample_from_nodes(node_idx)
+# node = out.node
+# edge = [out.row, out.col]
+
+end = time.time()
+print('node:', out.node)
+print('edge_index_list:', out.edge_index_list)
+print('eid_list:', out.eid_list)
+print('eid_ts_list:', out.eid_ts_list)
+print("sample time", end-pre)
\ No newline at end of file

starrygl/sample/sample_core/ogb_demo_static.py

+import torch
+from ogb.nodeproppred import PygNodePropPredDataset
+from torch_geometric import datasets
+import time
+from Utils import GraphData
+
+def load_ogb_dataset(name, data_path):
+    dataset = PygNodePropPredDataset(name=name, root=data_path)
+    split_idx = dataset.get_idx_split()
+    g = dataset[0]
+    n_node = g.num_nodes
+    node_data={}
+    node_data['train_mask'] = torch.zeros(n_node, dtype=torch.bool)
+    node_data['val_mask'] = torch.zeros(n_node, dtype=torch.bool)
+    node_data['test_mask'] = torch.zeros(n_node, dtype=torch.bool)
+    node_data['train_mask'][split_idx["train"]] = True
+    node_data['val_mask'][split_idx["valid"]] = True
+    node_data['test_mask'][split_idx["test"]] = True
+    return g, node_data
+
+g, node_data = load_ogb_dataset('ogbn-products', "/home/zlj/hzq/code/gnn/dataset/")
+print(g)
+# for worker in [1,2,3,4,5,6,7,8,9,10,20,30]:
+g_data = GraphData(id=1, edge_index=g.edge_index, data=g, partptr=torch.tensor([0, g.num_nodes//4, g.num_nodes//4*2, g.num_nodes//4*3, g.num_nodes]))
+
+row, col = g.edge_index
+
+# edge_weight = torch.ones(g.num_edges).float()
+# indices = [x for x in range(0, g.num_edges, 5)]
+# edge_weight[indices] = 2.0
+
+# g_data.eid = None
+edge_weight = None
+timestamp = None
+
+from neighbor_sampler import NeighborSampler, SampleType
+from neighbor_sampler import get_neighbors
+
+update_edge_row =     row
+update_edge_col =     col
+update_edge_w =  torch.DoubleTensor([i for i in range(g.num_edges)])
+print('begin update')
+pre = time.time()
+# update_edge_weight(tnb, update_edge_row.contiguous(), update_edge_col.contiguous(), update_edge_w.contiguous())
+end = time.time()
+print("update time:", end-pre)
+
+print('begin tnb')
+pre = time.time()
+tnb = get_neighbors("a",
+                    row.contiguous(), 
+                    col.contiguous(), 
+                    g.num_nodes, 0, 
+                    g_data.eid, 
+                    edge_weight, 
+                    timestamp)
+end = time.time()
+print("init tnb time:", end-pre)
+torch.save(tnb, "/home/zlj/hzq/code/gnn/my_sampler/MergeSample/tnb_static.my")
+
+
+# print('begin load')
+# pre = time.time()
+# tnb = torch.load("/home/zlj/hzq/code/gnn/my_sampler/MergeSample/tnb_static.my")
+# end = time.time()
+# print("load time:", end-pre)
+
+print('begin init')
+pre = time.time()
+sampler = NeighborSampler(g.num_nodes, 
+                          tnb = tnb, 
+                          num_layers=2, 
+                          fanout=[100,100], 
+                          graph_data=g_data, 
+                          workers=10, 
+                          graph_name='a',
+                          policy="uniform")
+end = time.time()
+print("init time:", end-pre)
+
+# from torch_geometric.sampler import NeighborSampler, NumNeighbors, NodeSamplerInput, SamplerOutput
+# pre = time.time()
+# num_nei = NumNeighbors([100, 100])
+# node_idx = NodeSamplerInput(input_id=None, node=torch.tensor(range(g.num_nodes//4, g.num_nodes//4+600000)))# (input_id=None, node=torch.masked_select(torch.arange(g.num_nodes),node_data['train_mask']))
+# sampler = NeighborSampler(g, num_nei)
+# end = time.time()
+# print("init time:", end-pre)
+
+pre = time.time()
+
+out = sampler.sample_from_nodes(torch.tensor(range(g.num_nodes//4, g.num_nodes//4+600000)), with_outer_sample=SampleType.Inner)# sampler.sample_from_nodes(torch.masked_select(torch.arange(g.num_nodes),node_data['train_mask']))
+
+# out = sampler.sample_from_nodes(node_idx)
+# node = out.node
+# edge = [out.row, out.col]
+
+end = time.time()
+print('node1:\t', out[0].sample_nodes())
+print('eid1:\t', out[0].eid())
+print('edge1:\t', g.edge_index[:, out[0].eid()])
+print('node2:\t', out[1].sample_nodes())
+print('eid2:\t', out[1].eid())
+print('edge2:\t', g.edge_index[:, out[1].eid()])
+print("sample time", end-pre)
\ No newline at end of file

starrygl/sample/sample_core/parallel_hashmap/conanfile.py

+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+from conans import ConanFile, tools
+import os
+
+class SparseppConan(ConanFile):
+    name = "parallel_hashmap"
+    version = "1.27"
+    description = "A header-only, very fast and memory-friendly hash map"
+    
+    # Indicates License type of the packaged library
+    license = "https://github.com/greg7mdp/parallel-hashmap/blob/master/LICENSE"
+    
+    # Packages the license for the conanfile.py
+    exports = ["LICENSE"]
+    
+    # Custom attributes for Bincrafters recipe conventions
+    source_subfolder = "source_subfolder"
+    
+    def source(self):
+        source_url = "https://github.com/greg7mdp/parallel-hashmap"
+        tools.get("{0}/archive/{1}.tar.gz".format(source_url, self.version))
+        extracted_dir = self.name + "-" + self.version
+
+        #Rename to "source_folder" is a convention to simplify later steps
+        os.rename(extracted_dir, self.source_subfolder)
+
+
+    def package(self):
+        include_folder = os.path.join(self.source_subfolder, "parallel_hashmap")
+        self.copy(pattern="LICENSE")
+        self.copy(pattern="*", dst="include/parallel_hashmap", src=include_folder)
+
+    def package_id(self):
+        self.info.header_only()

starrygl/sample/sample_core/parallel_hashmap/meminfo.h

+#if !defined(spp_memory_h_guard)
+#define spp_memory_h_guard
+
+#include <cstdint>
+#include <cstring>
+#include <cstdlib>
+
+#if defined(_WIN32) || defined( __CYGWIN__)
+    #define SPP_WIN
+#endif
+
+#ifdef SPP_WIN
+    #include <windows.h>
+    #include <Psapi.h>
+    #undef min
+    #undef max
+#elif defined(__linux__)
+    #include <sys/types.h>
+    #include <sys/sysinfo.h>
+#elif defined(__FreeBSD__)
+    #include <paths.h>
+    #include <fcntl.h>
+    #include <kvm.h>
+    #include <unistd.h>
+    #include <sys/sysctl.h>
+    #include <sys/user.h>
+#endif
+
+namespace spp
+{
+    uint64_t GetSystemMemory();
+    uint64_t GetTotalMemoryUsed();
+    uint64_t GetProcessMemoryUsed();
+    uint64_t GetPhysicalMemory();
+
+    uint64_t GetSystemMemory()
+    {
+#ifdef SPP_WIN
+        MEMORYSTATUSEX memInfo;
+        memInfo.dwLength = sizeof(MEMORYSTATUSEX);
+        GlobalMemoryStatusEx(&memInfo);
+        return static_cast<uint64_t>(memInfo.ullTotalPageFile);
+#elif defined(__linux__)
+        struct sysinfo memInfo;
+        sysinfo (&memInfo);
+        auto totalVirtualMem = memInfo.totalram;
+
+        totalVirtualMem += memInfo.totalswap;
+        totalVirtualMem *= memInfo.mem_unit;
+        return static_cast<uint64_t>(totalVirtualMem);
+#elif defined(__FreeBSD__)
+        kvm_t *kd;
+        u_int pageCnt;
+        size_t pageCntLen = sizeof(pageCnt);
+        u_int pageSize;
+        struct kvm_swap kswap;
+        uint64_t totalVirtualMem;
+
+        pageSize = static_cast<u_int>(getpagesize());
+
+        sysctlbyname("vm.stats.vm.v_page_count", &pageCnt, &pageCntLen, NULL, 0);
+        totalVirtualMem = pageCnt * pageSize;
+
+        kd = kvm_open(NULL, _PATH_DEVNULL, NULL, O_RDONLY, "kvm_open");
+        kvm_getswapinfo(kd, &kswap, 1, 0);
+        kvm_close(kd);
+        totalVirtualMem += kswap.ksw_total * pageSize;
+
+        return totalVirtualMem;
+#else
+        return 0;
+#endif
+    }
+
+    uint64_t GetTotalMemoryUsed()
+    {
+#ifdef SPP_WIN
+        MEMORYSTATUSEX memInfo;
+        memInfo.dwLength = sizeof(MEMORYSTATUSEX);
+        GlobalMemoryStatusEx(&memInfo);
+        return static_cast<uint64_t>(memInfo.ullTotalPageFile - memInfo.ullAvailPageFile);
+#elif defined(__linux__)
+        struct sysinfo memInfo;
+        sysinfo(&memInfo);
+        auto virtualMemUsed = memInfo.totalram - memInfo.freeram;
+
+        virtualMemUsed += memInfo.totalswap - memInfo.freeswap;
+        virtualMemUsed *= memInfo.mem_unit;
+
+        return static_cast<uint64_t>(virtualMemUsed);
+#elif defined(__FreeBSD__)
+        kvm_t *kd;
+        u_int pageSize;
+        u_int pageCnt, freeCnt;
+        size_t pageCntLen = sizeof(pageCnt);
+        size_t freeCntLen = sizeof(freeCnt);
+        struct kvm_swap kswap;
+        uint64_t virtualMemUsed;
+
+        pageSize = static_cast<u_int>(getpagesize());
+
+        sysctlbyname("vm.stats.vm.v_page_count", &pageCnt, &pageCntLen, NULL, 0);
+        sysctlbyname("vm.stats.vm.v_free_count", &freeCnt, &freeCntLen, NULL, 0);
+        virtualMemUsed = (pageCnt - freeCnt) * pageSize;
+
+        kd = kvm_open(NULL, _PATH_DEVNULL, NULL, O_RDONLY, "kvm_open");
+        kvm_getswapinfo(kd, &kswap, 1, 0);
+        kvm_close(kd);
+        virtualMemUsed += kswap.ksw_used * pageSize;
+
+        return virtualMemUsed;
+#else
+        return 0;
+#endif
+    }
+
+    uint64_t GetProcessMemoryUsed()
+    {
+#ifdef SPP_WIN
+        PROCESS_MEMORY_COUNTERS_EX pmc;
+        GetProcessMemoryInfo(GetCurrentProcess(), reinterpret_cast<PPROCESS_MEMORY_COUNTERS>(&pmc), sizeof(pmc));
+        return static_cast<uint64_t>(pmc.PrivateUsage);
+#elif defined(__linux__)
+        auto parseLine = 
+            [](char* line)->int
+            {
+                auto i = strlen(line);
+				
+                while(*line < '0' || *line > '9') 
+                {
+                    line++;
+                }
+
+                line[i-3] = '\0';
+                i = atoi(line);
+                return i;
+            };
+
+        auto file = fopen("/proc/self/status", "r");
+        auto result = -1;
+        char line[128];
+
+        while(fgets(line, 128, file) != nullptr)
+        {
+            if(strncmp(line, "VmSize:", 7) == 0)
+            {
+                result = parseLine(line);
+                break;
+            }
+        }
+
+        fclose(file);
+        return static_cast<uint64_t>(result) * 1024;
+#elif defined(__FreeBSD__)
+        struct kinfo_proc info;
+        size_t infoLen = sizeof(info);
+        int mib[] = { CTL_KERN, KERN_PROC, KERN_PROC_PID, getpid() };
+
+        sysctl(mib, sizeof(mib) / sizeof(*mib), &info, &infoLen, NULL, 0);
+        return static_cast<uint64_t>(info.ki_rssize * getpagesize());
+#else
+        return 0;
+#endif
+    }
+
+    uint64_t GetPhysicalMemory()
+    {
+#ifdef SPP_WIN
+        MEMORYSTATUSEX memInfo;
+        memInfo.dwLength = sizeof(MEMORYSTATUSEX);
+        GlobalMemoryStatusEx(&memInfo);
+        return static_cast<uint64_t>(memInfo.ullTotalPhys);
+#elif defined(__linux__)
+        struct sysinfo memInfo;
+        sysinfo(&memInfo);
+
+        auto totalPhysMem = memInfo.totalram;
+
+        totalPhysMem *= memInfo.mem_unit;
+        return static_cast<uint64_t>(totalPhysMem);
+#elif defined(__FreeBSD__)
+        u_long physMem;
+        size_t physMemLen = sizeof(physMem);
+        int mib[] = { CTL_HW, HW_PHYSMEM };
+
+        sysctl(mib, sizeof(mib) / sizeof(*mib), &physMem, &physMemLen, NULL, 0);
+        return physMem;
+#else
+        return 0;
+#endif
+    }
+
+}
+
+#endif // spp_memory_h_guard

starrygl/sample/sample_core/parallel_hashmap/phmap_bits.h

+#if !defined(phmap_bits_h_guard_)
+#define phmap_bits_h_guard_
+
+// ---------------------------------------------------------------------------
+// Copyright (c) 2019, Gregory Popovitch - greg7mdp@gmail.com
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//
+// Includes work from abseil-cpp (https://github.com/abseil/abseil-cpp)
+// with modifications.
+// 
+// Copyright 2018 The Abseil Authors.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+// ---------------------------------------------------------------------------
+
+// The following guarantees declaration of the byte swap functions
+#ifdef _MSC_VER
+    #include <stdlib.h>  // NOLINT(build/include)
+#elif defined(__APPLE__)
+    // Mac OS X / Darwin features
+    #include <libkern/OSByteOrder.h>
+#elif defined(__FreeBSD__)
+    #include <sys/endian.h>
+#elif defined(__GLIBC__)
+    #include <byteswap.h>  // IWYU pragma: export
+#endif
+
+#include <string.h>
+#include <cstdint>
+#include "phmap_config.h"
+
+#ifdef _MSC_VER
+    #pragma warning(push)  
+    #pragma warning(disable : 4514) // unreferenced inline function has been removed
+#endif
+
+// -----------------------------------------------------------------------------
+// unaligned APIs
+// -----------------------------------------------------------------------------
+// Portable handling of unaligned loads, stores, and copies.
+// On some platforms, like ARM, the copy functions can be more efficient
+// then a load and a store.
+// -----------------------------------------------------------------------------
+
+#if defined(ADDRESS_SANITIZER) || defined(THREAD_SANITIZER) ||\
+    defined(MEMORY_SANITIZER)
+#include <stdint.h>
+
+extern "C" {
+    uint16_t __sanitizer_unaligned_load16(const void *p);
+    uint32_t __sanitizer_unaligned_load32(const void *p);
+    uint64_t __sanitizer_unaligned_load64(const void *p);
+    void __sanitizer_unaligned_store16(void *p, uint16_t v);
+    void __sanitizer_unaligned_store32(void *p, uint32_t v);
+    void __sanitizer_unaligned_store64(void *p, uint64_t v);
+}  // extern "C"
+
+namespace phmap {
+namespace bits {
+
+inline uint16_t UnalignedLoad16(const void *p) {
+  return __sanitizer_unaligned_load16(p);
+}
+
+inline uint32_t UnalignedLoad32(const void *p) {
+  return __sanitizer_unaligned_load32(p);
+}
+
+inline uint64_t UnalignedLoad64(const void *p) {
+  return __sanitizer_unaligned_load64(p);
+}
+
+inline void UnalignedStore16(void *p, uint16_t v) {
+  __sanitizer_unaligned_store16(p, v);
+}
+
+inline void UnalignedStore32(void *p, uint32_t v) {
+  __sanitizer_unaligned_store32(p, v);
+}
+
+inline void UnalignedStore64(void *p, uint64_t v) {
+  __sanitizer_unaligned_store64(p, v);
+}
+
+}  // namespace bits
+}  // namespace phmap
+
+#define PHMAP_INTERNAL_UNALIGNED_LOAD16(_p) (phmap::bits::UnalignedLoad16(_p))
+#define PHMAP_INTERNAL_UNALIGNED_LOAD32(_p) (phmap::bits::UnalignedLoad32(_p))
+#define PHMAP_INTERNAL_UNALIGNED_LOAD64(_p) (phmap::bits::UnalignedLoad64(_p))
+
+#define PHMAP_INTERNAL_UNALIGNED_STORE16(_p, _val) (phmap::bits::UnalignedStore16(_p, _val))
+#define PHMAP_INTERNAL_UNALIGNED_STORE32(_p, _val) (phmap::bits::UnalignedStore32(_p, _val))
+#define PHMAP_INTERNAL_UNALIGNED_STORE64(_p, _val) (phmap::bits::UnalignedStore64(_p, _val))
+
+#else
+
+namespace phmap {
+namespace bits {
+
+inline uint16_t UnalignedLoad16(const void *p) {
+  uint16_t t;
+  memcpy(&t, p, sizeof t);
+  return t;
+}
+
+inline uint32_t UnalignedLoad32(const void *p) {
+  uint32_t t;
+  memcpy(&t, p, sizeof t);
+  return t;
+}
+
+inline uint64_t UnalignedLoad64(const void *p) {
+  uint64_t t;
+  memcpy(&t, p, sizeof t);
+  return t;
+}
+
+inline void UnalignedStore16(void *p, uint16_t v) { memcpy(p, &v, sizeof v); }
+
+inline void UnalignedStore32(void *p, uint32_t v) { memcpy(p, &v, sizeof v); }
+
+inline void UnalignedStore64(void *p, uint64_t v) { memcpy(p, &v, sizeof v); }
+
+}  // namespace bits
+}  // namespace phmap
+
+#define PHMAP_INTERNAL_UNALIGNED_LOAD16(_p) (phmap::bits::UnalignedLoad16(_p))
+#define PHMAP_INTERNAL_UNALIGNED_LOAD32(_p) (phmap::bits::UnalignedLoad32(_p))
+#define PHMAP_INTERNAL_UNALIGNED_LOAD64(_p) (phmap::bits::UnalignedLoad64(_p))
+
+#define PHMAP_INTERNAL_UNALIGNED_STORE16(_p, _val) (phmap::bits::UnalignedStore16(_p, _val))
+#define PHMAP_INTERNAL_UNALIGNED_STORE32(_p, _val) (phmap::bits::UnalignedStore32(_p, _val))
+#define PHMAP_INTERNAL_UNALIGNED_STORE64(_p, _val) (phmap::bits::UnalignedStore64(_p, _val))
+
+#endif
+
+// -----------------------------------------------------------------------------
+// File: optimization.h
+// -----------------------------------------------------------------------------
+
+#if defined(__pnacl__)
+    #define PHMAP_BLOCK_TAIL_CALL_OPTIMIZATION() if (volatile int x = 0) { (void)x; }
+#elif defined(__clang__)
+    // Clang will not tail call given inline volatile assembly.
+    #define PHMAP_BLOCK_TAIL_CALL_OPTIMIZATION() __asm__ __volatile__("")
+#elif defined(__GNUC__)
+    // GCC will not tail call given inline volatile assembly.
+    #define PHMAP_BLOCK_TAIL_CALL_OPTIMIZATION() __asm__ __volatile__("")
+#elif defined(_MSC_VER)
+    #include <intrin.h>
+    // The __nop() intrinsic blocks the optimisation.
+    #define PHMAP_BLOCK_TAIL_CALL_OPTIMIZATION() __nop()
+#else
+    #define PHMAP_BLOCK_TAIL_CALL_OPTIMIZATION() if (volatile int x = 0) { (void)x; }
+#endif
+
+#if defined(__GNUC__)
+    #pragma GCC diagnostic push
+    #pragma GCC diagnostic ignored "-Wpedantic"
+#endif
+
+#ifdef PHMAP_HAVE_INTRINSIC_INT128
+    __extension__ typedef unsigned __int128 phmap_uint128;
+    inline uint64_t umul128(uint64_t a, uint64_t b, uint64_t* high) 
+    {
+        auto result = static_cast<phmap_uint128>(a) * static_cast<phmap_uint128>(b);
+        *high = static_cast<uint64_t>(result >> 64);
+        return static_cast<uint64_t>(result);
+    }
+    #define PHMAP_HAS_UMUL128 1
+#elif (defined(_MSC_VER))
+    #if defined(_M_X64)
+        #pragma intrinsic(_umul128)
+        inline uint64_t umul128(uint64_t a, uint64_t b, uint64_t* high) 
+        {
+            return _umul128(a, b, high);
+        }
+        #define PHMAP_HAS_UMUL128 1
+    #endif
+#endif
+
+#if defined(__GNUC__)
+    #pragma GCC diagnostic pop
+#endif
+
+#if defined(__GNUC__)
+    // Cache line alignment
+    #if defined(__i386__) || defined(__x86_64__)
+        #define PHMAP_CACHELINE_SIZE 64
+    #elif defined(__powerpc64__)
+        #define PHMAP_CACHELINE_SIZE 128
+    #elif defined(__aarch64__)
+        // We would need to read special register ctr_el0 to find out L1 dcache size.
+        // This value is a good estimate based on a real aarch64 machine.
+        #define PHMAP_CACHELINE_SIZE 64
+    #elif defined(__arm__)
+        // Cache line sizes for ARM: These values are not strictly correct since
+        // cache line sizes depend on implementations, not architectures.  There
+        // are even implementations with cache line sizes configurable at boot
+        // time.
+        #if defined(__ARM_ARCH_5T__)
+            #define PHMAP_CACHELINE_SIZE 32
+        #elif defined(__ARM_ARCH_7A__)
+            #define PHMAP_CACHELINE_SIZE 64
+        #endif
+    #endif
+
+    #ifndef PHMAP_CACHELINE_SIZE
+        // A reasonable default guess.  Note that overestimates tend to waste more
+        // space, while underestimates tend to waste more time.
+        #define PHMAP_CACHELINE_SIZE 64
+    #endif
+
+    #define PHMAP_CACHELINE_ALIGNED __attribute__((aligned(PHMAP_CACHELINE_SIZE)))
+#elif defined(_MSC_VER)
+    #define PHMAP_CACHELINE_SIZE 64
+    #define PHMAP_CACHELINE_ALIGNED __declspec(align(PHMAP_CACHELINE_SIZE))
+#else
+    #define PHMAP_CACHELINE_SIZE 64
+    #define PHMAP_CACHELINE_ALIGNED
+#endif
+
+
+#if PHMAP_HAVE_BUILTIN(__builtin_expect) || \
+    (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_PREDICT_FALSE(x) (__builtin_expect(x, 0))
+    #define PHMAP_PREDICT_TRUE(x) (__builtin_expect(!!(x), 1))
+#else
+    #define PHMAP_PREDICT_FALSE(x) (x)
+    #define PHMAP_PREDICT_TRUE(x) (x)
+#endif
+
+// -----------------------------------------------------------------------------
+// File: bits.h
+// -----------------------------------------------------------------------------
+
+#if defined(_MSC_VER)
+    // We can achieve something similar to attribute((always_inline)) with MSVC by
+    // using the __forceinline keyword, however this is not perfect. MSVC is
+    // much less aggressive about inlining, and even with the __forceinline keyword.
+    #define PHMAP_BASE_INTERNAL_FORCEINLINE __forceinline
+#else
+    // Use default attribute inline.
+    #define PHMAP_BASE_INTERNAL_FORCEINLINE inline PHMAP_ATTRIBUTE_ALWAYS_INLINE
+#endif
+
+
+namespace phmap {
+namespace base_internal {
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountLeadingZeros64Slow(uint64_t n) {
+    int zeroes = 60;
+    if (n >> 32) zeroes -= 32, n >>= 32;
+    if (n >> 16) zeroes -= 16, n >>= 16;
+    if (n >> 8) zeroes -= 8, n >>= 8;
+    if (n >> 4) zeroes -= 4, n >>= 4;
+    return "\4\3\2\2\1\1\1\1\0\0\0\0\0\0\0"[n] + zeroes;
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountLeadingZeros64(uint64_t n) {
+#if defined(_MSC_VER) && defined(_M_X64)
+    // MSVC does not have __buitin_clzll. Use _BitScanReverse64.
+    unsigned long result = 0;  // NOLINT(runtime/int)
+    if (_BitScanReverse64(&result, n)) {
+        return (int)(63 - result);
+    }
+    return 64;
+#elif defined(_MSC_VER) && !defined(__clang__)
+    // MSVC does not have __buitin_clzll. Compose two calls to _BitScanReverse
+    unsigned long result = 0;  // NOLINT(runtime/int)
+    if ((n >> 32) && _BitScanReverse(&result, (unsigned long)(n >> 32))) {
+        return 31 - result;
+    }
+    if (_BitScanReverse(&result, (unsigned long)n)) {
+        return 63 - result;
+    }
+    return 64;
+#elif defined(__GNUC__) || defined(__clang__)
+    // Use __builtin_clzll, which uses the following instructions:
+    //  x86: bsr
+    //  ARM64: clz
+    //  PPC: cntlzd
+    static_assert(sizeof(unsigned long long) == sizeof(n),  // NOLINT(runtime/int)
+                  "__builtin_clzll does not take 64-bit arg");
+
+    // Handle 0 as a special case because __builtin_clzll(0) is undefined.
+    if (n == 0) {
+        return 64;
+    }
+    return __builtin_clzll(n);
+#else
+    return CountLeadingZeros64Slow(n);
+#endif
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountLeadingZeros32Slow(uint64_t n) {
+    int zeroes = 28;
+    if (n >> 16) zeroes -= 16, n >>= 16;
+    if (n >> 8) zeroes -= 8, n >>= 8;
+    if (n >> 4) zeroes -= 4, n >>= 4;
+    return "\4\3\2\2\1\1\1\1\0\0\0\0\0\0\0"[n] + zeroes;
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountLeadingZeros32(uint32_t n) {
+#if defined(_MSC_VER) && !defined(__clang__)
+    unsigned long result = 0;  // NOLINT(runtime/int)
+    if (_BitScanReverse(&result, n)) {
+        return (int)(31 - result);
+    }
+    return 32;
+#elif defined(__GNUC__) || defined(__clang__)
+    // Use __builtin_clz, which uses the following instructions:
+    //  x86: bsr
+    //  ARM64: clz
+    //  PPC: cntlzd
+    static_assert(sizeof(int) == sizeof(n),
+                  "__builtin_clz does not take 32-bit arg");
+
+    // Handle 0 as a special case because __builtin_clz(0) is undefined.
+    if (n == 0) {
+        return 32;
+    }
+    return __builtin_clz(n);
+#else
+    return CountLeadingZeros32Slow(n);
+#endif
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountTrailingZerosNonZero64Slow(uint64_t n) {
+    int c = 63;
+    n &= ~n + 1;
+    if (n & 0x00000000FFFFFFFF) c -= 32;
+    if (n & 0x0000FFFF0000FFFF) c -= 16;
+    if (n & 0x00FF00FF00FF00FF) c -= 8;
+    if (n & 0x0F0F0F0F0F0F0F0F) c -= 4;
+    if (n & 0x3333333333333333) c -= 2;
+    if (n & 0x5555555555555555) c -= 1;
+    return c;
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountTrailingZerosNonZero64(uint64_t n) {
+#if defined(_MSC_VER) && !defined(__clang__) && defined(_M_X64)
+    unsigned long result = 0;  // NOLINT(runtime/int)
+    _BitScanForward64(&result, n);
+    return (int)result;
+#elif defined(_MSC_VER) && !defined(__clang__)
+    unsigned long result = 0;  // NOLINT(runtime/int)
+    if (static_cast<uint32_t>(n) == 0) {
+        _BitScanForward(&result, (unsigned long)(n >> 32));
+        return result + 32;
+    }
+    _BitScanForward(&result, (unsigned long)n);
+    return result;
+#elif defined(__GNUC__) || defined(__clang__)
+    static_assert(sizeof(unsigned long long) == sizeof(n),  // NOLINT(runtime/int)
+                  "__builtin_ctzll does not take 64-bit arg");
+    return __builtin_ctzll(n);
+#else
+    return CountTrailingZerosNonZero64Slow(n);
+#endif
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountTrailingZerosNonZero32Slow(uint32_t n) {
+    int c = 31;
+    n &= ~n + 1;
+    if (n & 0x0000FFFF) c -= 16;
+    if (n & 0x00FF00FF) c -= 8;
+    if (n & 0x0F0F0F0F) c -= 4;
+    if (n & 0x33333333) c -= 2;
+    if (n & 0x55555555) c -= 1;
+    return c;
+}
+
+PHMAP_BASE_INTERNAL_FORCEINLINE int CountTrailingZerosNonZero32(uint32_t n) {
+#if defined(_MSC_VER) && !defined(__clang__)
+    unsigned long result = 0;  // NOLINT(runtime/int)
+    _BitScanForward(&result, n);
+    return (int)result;
+#elif defined(__GNUC__) || defined(__clang__)
+    static_assert(sizeof(int) == sizeof(n),
+                  "__builtin_ctz does not take 32-bit arg");
+    return __builtin_ctz(n);
+#else
+    return CountTrailingZerosNonZero32Slow(n);
+#endif
+}
+
+#undef PHMAP_BASE_INTERNAL_FORCEINLINE
+
+}  // namespace base_internal
+}  // namespace phmap
+
+// -----------------------------------------------------------------------------
+// File: endian.h
+// -----------------------------------------------------------------------------
+
+namespace phmap {
+
+// Use compiler byte-swapping intrinsics if they are available.  32-bit
+// and 64-bit versions are available in Clang and GCC as of GCC 4.3.0.
+// The 16-bit version is available in Clang and GCC only as of GCC 4.8.0.
+// For simplicity, we enable them all only for GCC 4.8.0 or later.
+#if defined(__clang__) || \
+    (defined(__GNUC__) && \
+     ((__GNUC__ == 4 && __GNUC_MINOR__ >= 8) || __GNUC__ >= 5))
+
+    inline uint64_t gbswap_64(uint64_t host_int) {
+        return __builtin_bswap64(host_int);
+    }
+    inline uint32_t gbswap_32(uint32_t host_int) {
+        return __builtin_bswap32(host_int);
+    }
+    inline uint16_t gbswap_16(uint16_t host_int) {
+        return __builtin_bswap16(host_int);
+    }
+
+#elif defined(_MSC_VER)
+
+    inline uint64_t gbswap_64(uint64_t host_int) {
+        return _byteswap_uint64(host_int);
+    }
+    inline uint32_t gbswap_32(uint32_t host_int) {
+        return _byteswap_ulong(host_int);
+    }
+    inline uint16_t gbswap_16(uint16_t host_int) {
+        return _byteswap_ushort(host_int);
+    }
+
+#elif defined(__APPLE__)
+
+    inline uint64_t gbswap_64(uint64_t host_int) { return OSSwapInt16(host_int); }
+    inline uint32_t gbswap_32(uint32_t host_int) { return OSSwapInt32(host_int); }
+    inline uint16_t gbswap_16(uint16_t host_int) { return OSSwapInt64(host_int); }
+
+#else
+
+    inline uint64_t gbswap_64(uint64_t host_int) {
+#if defined(__GNUC__) && defined(__x86_64__) && !defined(__APPLE__)
+        // Adapted from /usr/include/byteswap.h.  Not available on Mac.
+        if (__builtin_constant_p(host_int)) {
+            return __bswap_constant_64(host_int);
+        } else {
+            uint64_t result;
+            __asm__("bswap %0" : "=r"(result) : "0"(host_int));
+            return result;
+        }
+#elif defined(__GLIBC__)
+        return bswap_64(host_int);
+#else
+        return (((host_int & uint64_t{0xFF}) << 56) |
+                ((host_int & uint64_t{0xFF00}) << 40) |
+                ((host_int & uint64_t{0xFF0000}) << 24) |
+                ((host_int & uint64_t{0xFF000000}) << 8) |
+                ((host_int & uint64_t{0xFF00000000}) >> 8) |
+                ((host_int & uint64_t{0xFF0000000000}) >> 24) |
+                ((host_int & uint64_t{0xFF000000000000}) >> 40) |
+                ((host_int & uint64_t{0xFF00000000000000}) >> 56));
+#endif  // bswap_64
+    }
+
+    inline uint32_t gbswap_32(uint32_t host_int) {
+#if defined(__GLIBC__)
+        return bswap_32(host_int);
+#else
+        return (((host_int & uint32_t{0xFF}) << 24) |
+                ((host_int & uint32_t{0xFF00}) << 8) |
+                ((host_int & uint32_t{0xFF0000}) >> 8) |
+                ((host_int & uint32_t{0xFF000000}) >> 24));
+#endif
+    }
+
+    inline uint16_t gbswap_16(uint16_t host_int) {
+#if defined(__GLIBC__)
+        return bswap_16(host_int);
+#else
+        return (((host_int & uint16_t{0xFF}) << 8) |
+                ((host_int & uint16_t{0xFF00}) >> 8));
+#endif
+    }
+
+#endif  // intrinics available
+
+#ifdef PHMAP_IS_LITTLE_ENDIAN
+
+    // Definitions for ntohl etc. that don't require us to include
+    // netinet/in.h. We wrap gbswap_32 and gbswap_16 in functions rather
+    // than just #defining them because in debug mode, gcc doesn't
+    // correctly handle the (rather involved) definitions of bswap_32.
+    // gcc guarantees that inline functions are as fast as macros, so
+    // this isn't a performance hit.
+    inline uint16_t ghtons(uint16_t x) { return gbswap_16(x); }
+    inline uint32_t ghtonl(uint32_t x) { return gbswap_32(x); }
+    inline uint64_t ghtonll(uint64_t x) { return gbswap_64(x); }
+
+#elif defined PHMAP_IS_BIG_ENDIAN
+
+    // These definitions are simpler on big-endian machines
+    // These are functions instead of macros to avoid self-assignment warnings
+    // on calls such as "i = ghtnol(i);".  This also provides type checking.
+    inline uint16_t ghtons(uint16_t x) { return x; }
+    inline uint32_t ghtonl(uint32_t x) { return x; }
+    inline uint64_t ghtonll(uint64_t x) { return x; }
+
+#else
+    #error \
+        "Unsupported byte order: Either PHMAP_IS_BIG_ENDIAN or " \
+           "PHMAP_IS_LITTLE_ENDIAN must be defined"
+#endif  // byte order
+
+inline uint16_t gntohs(uint16_t x) { return ghtons(x); }
+inline uint32_t gntohl(uint32_t x) { return ghtonl(x); }
+inline uint64_t gntohll(uint64_t x) { return ghtonll(x); }
+
+// Utilities to convert numbers between the current hosts's native byte
+// order and little-endian byte order
+//
+// Load/Store methods are alignment safe
+namespace little_endian {
+// Conversion functions.
+#ifdef PHMAP_IS_LITTLE_ENDIAN
+
+    inline uint16_t FromHost16(uint16_t x) { return x; }
+    inline uint16_t ToHost16(uint16_t x) { return x; }
+
+    inline uint32_t FromHost32(uint32_t x) { return x; }
+    inline uint32_t ToHost32(uint32_t x) { return x; }
+
+    inline uint64_t FromHost64(uint64_t x) { return x; }
+    inline uint64_t ToHost64(uint64_t x) { return x; }
+
+    inline constexpr bool IsLittleEndian() { return true; }
+
+#elif defined PHMAP_IS_BIG_ENDIAN
+
+    inline uint16_t FromHost16(uint16_t x) { return gbswap_16(x); }
+    inline uint16_t ToHost16(uint16_t x) { return gbswap_16(x); }
+
+    inline uint32_t FromHost32(uint32_t x) { return gbswap_32(x); }
+    inline uint32_t ToHost32(uint32_t x) { return gbswap_32(x); }
+
+    inline uint64_t FromHost64(uint64_t x) { return gbswap_64(x); }
+    inline uint64_t ToHost64(uint64_t x) { return gbswap_64(x); }
+
+    inline constexpr bool IsLittleEndian() { return false; }
+
+#endif /* ENDIAN */
+
+// Functions to do unaligned loads and stores in little-endian order.
+inline uint16_t Load16(const void *p) {
+  return ToHost16(PHMAP_INTERNAL_UNALIGNED_LOAD16(p));
+}
+
+inline void Store16(void *p, uint16_t v) {
+  PHMAP_INTERNAL_UNALIGNED_STORE16(p, FromHost16(v));
+}
+
+inline uint32_t Load32(const void *p) {
+  return ToHost32(PHMAP_INTERNAL_UNALIGNED_LOAD32(p));
+}
+
+inline void Store32(void *p, uint32_t v) {
+  PHMAP_INTERNAL_UNALIGNED_STORE32(p, FromHost32(v));
+}
+
+inline uint64_t Load64(const void *p) {
+  return ToHost64(PHMAP_INTERNAL_UNALIGNED_LOAD64(p));
+}
+
+inline void Store64(void *p, uint64_t v) {
+  PHMAP_INTERNAL_UNALIGNED_STORE64(p, FromHost64(v));
+}
+
+}  // namespace little_endian
+
+// Utilities to convert numbers between the current hosts's native byte
+// order and big-endian byte order (same as network byte order)
+//
+// Load/Store methods are alignment safe
+namespace big_endian {
+#ifdef PHMAP_IS_LITTLE_ENDIAN
+
+    inline uint16_t FromHost16(uint16_t x) { return gbswap_16(x); }
+    inline uint16_t ToHost16(uint16_t x) { return gbswap_16(x); }
+
+    inline uint32_t FromHost32(uint32_t x) { return gbswap_32(x); }
+    inline uint32_t ToHost32(uint32_t x) { return gbswap_32(x); }
+
+    inline uint64_t FromHost64(uint64_t x) { return gbswap_64(x); }
+    inline uint64_t ToHost64(uint64_t x) { return gbswap_64(x); }
+
+    inline constexpr bool IsLittleEndian() { return true; }
+
+#elif defined PHMAP_IS_BIG_ENDIAN
+
+    inline uint16_t FromHost16(uint16_t x) { return x; }
+    inline uint16_t ToHost16(uint16_t x) { return x; }
+
+    inline uint32_t FromHost32(uint32_t x) { return x; }
+    inline uint32_t ToHost32(uint32_t x) { return x; }
+
+    inline uint64_t FromHost64(uint64_t x) { return x; }
+    inline uint64_t ToHost64(uint64_t x) { return x; }
+
+    inline constexpr bool IsLittleEndian() { return false; }
+
+#endif /* ENDIAN */
+
+// Functions to do unaligned loads and stores in big-endian order.
+inline uint16_t Load16(const void *p) {
+  return ToHost16(PHMAP_INTERNAL_UNALIGNED_LOAD16(p));
+}
+
+inline void Store16(void *p, uint16_t v) {
+  PHMAP_INTERNAL_UNALIGNED_STORE16(p, FromHost16(v));
+}
+
+inline uint32_t Load32(const void *p) {
+  return ToHost32(PHMAP_INTERNAL_UNALIGNED_LOAD32(p));
+}
+
+inline void Store32(void *p, uint32_t v) {
+  PHMAP_INTERNAL_UNALIGNED_STORE32(p, FromHost32(v));
+}
+
+inline uint64_t Load64(const void *p) {
+  return ToHost64(PHMAP_INTERNAL_UNALIGNED_LOAD64(p));
+}
+
+inline void Store64(void *p, uint64_t v) {
+  PHMAP_INTERNAL_UNALIGNED_STORE64(p, FromHost64(v));
+}
+
+}  // namespace big_endian
+
+}  // namespace phmap
+
+#ifdef _MSC_VER
+     #pragma warning(pop)  
+#endif
+
+#endif // phmap_bits_h_guard_

starrygl/sample/sample_core/parallel_hashmap/phmap_config.h

+#if !defined(phmap_config_h_guard_)
+#define phmap_config_h_guard_
+
+// ---------------------------------------------------------------------------
+// Copyright (c) 2019, Gregory Popovitch - greg7mdp@gmail.com
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+//
+// Includes work from abseil-cpp (https://github.com/abseil/abseil-cpp)
+// with modifications.
+// 
+// Copyright 2018 The Abseil Authors.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+// ---------------------------------------------------------------------------
+
+#define PHMAP_VERSION_MAJOR 1
+#define PHMAP_VERSION_MINOR 0
+#define PHMAP_VERSION_PATCH 0
+
+// Included for the __GLIBC__ macro (or similar macros on other systems).
+#include <limits.h>
+
+#ifdef __cplusplus
+    // Included for __GLIBCXX__, _LIBCPP_VERSION
+    #include <cstddef>
+#endif  // __cplusplus
+
+#if defined(__APPLE__)
+    // Included for TARGET_OS_IPHONE, __IPHONE_OS_VERSION_MIN_REQUIRED,
+    // __IPHONE_8_0.
+    #include <Availability.h>
+    #include <TargetConditionals.h>
+#endif
+
+#define PHMAP_XSTR(x) PHMAP_STR(x)
+#define PHMAP_STR(x) #x
+#define PHMAP_VAR_NAME_VALUE(var) #var "="  PHMAP_STR(var)
+
+// -----------------------------------------------------------------------------
+// Some sanity checks
+// -----------------------------------------------------------------------------
+//#if defined(__CYGWIN__)
+//    #error "Cygwin is not supported."
+//#endif
+
+#if defined(_MSC_FULL_VER) && _MSC_FULL_VER < 190023918 && !defined(__clang__)
+    #error "phmap requires Visual Studio 2015 Update 2 or higher."
+#endif
+
+// We support gcc 4.7 and later.
+#if defined(__GNUC__) && !defined(__clang__)
+    #if __GNUC__ < 4 || (__GNUC__ == 4 && __GNUC_MINOR__ < 7)
+        #error "phmap requires gcc 4.7 or higher."
+    #endif
+#endif
+
+// We support Apple Xcode clang 4.2.1 (version 421.11.65) and later.
+// This corresponds to Apple Xcode version 4.5.
+#if defined(__apple_build_version__) && __apple_build_version__ < 4211165
+    #error "phmap requires __apple_build_version__ of 4211165 or higher."
+#endif
+
+// Enforce C++11 as the minimum. 
+#if defined(__cplusplus) && !defined(_MSC_VER)
+    #if __cplusplus < 201103L
+        #error "C++ versions less than C++11 are not supported."
+    #endif
+#endif
+
+// We have chosen glibc 2.12 as the minimum 
+#if defined(__GLIBC__) && defined(__GLIBC_PREREQ)
+    #if !__GLIBC_PREREQ(2, 12)
+        #error "Minimum required version of glibc is 2.12."
+    #endif
+#endif
+
+#if defined(_STLPORT_VERSION)
+    #error "STLPort is not supported."
+#endif
+
+#if CHAR_BIT != 8
+    #error "phmap assumes CHAR_BIT == 8."
+#endif
+
+// phmap currently assumes that an int is 4 bytes. 
+#if INT_MAX < 2147483647
+    #error "phmap assumes that int is at least 4 bytes. "
+#endif
+
+
+
+// -----------------------------------------------------------------------------
+// Compiler Feature Checks
+// -----------------------------------------------------------------------------
+
+#ifdef __has_builtin
+    #define PHMAP_HAVE_BUILTIN(x) __has_builtin(x)
+#else
+    #define PHMAP_HAVE_BUILTIN(x) 0
+#endif
+
+#if (defined(_MSVC_LANG) && _MSVC_LANG >= 201703) || __cplusplus >= 201703
+    #define PHMAP_HAVE_CC17 1
+#else
+    #define PHMAP_HAVE_CC17 0
+#endif
+
+#define PHMAP_BRANCHLESS 1
+
+// ----------------------------------------------------------------
+// Checks whether `std::is_trivially_destructible<T>` is supported.
+// ----------------------------------------------------------------
+#ifdef PHMAP_HAVE_STD_IS_TRIVIALLY_DESTRUCTIBLE
+    #error PHMAP_HAVE_STD_IS_TRIVIALLY_DESTRUCTIBLE cannot be directly set
+#elif defined(_LIBCPP_VERSION) ||                                        \
+    (!defined(__clang__) && defined(__GNUC__) && defined(__GLIBCXX__) && \
+     (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8))) ||        \
+    defined(_MSC_VER)
+    #define PHMAP_HAVE_STD_IS_TRIVIALLY_DESTRUCTIBLE 1
+#endif
+
+// --------------------------------------------------------------
+// Checks whether `std::is_trivially_default_constructible<T>` is 
+// supported.
+// --------------------------------------------------------------
+#if defined(PHMAP_HAVE_STD_IS_TRIVIALLY_CONSTRUCTIBLE)
+    #error PHMAP_HAVE_STD_IS_TRIVIALLY_CONSTRUCTIBLE cannot be directly set
+#elif defined(PHMAP_HAVE_STD_IS_TRIVIALLY_ASSIGNABLE)
+    #error PHMAP_HAVE_STD_IS_TRIVIALLY_ASSIGNABLE cannot directly set
+#elif (defined(__clang__) && defined(_LIBCPP_VERSION)) ||        \
+    (!defined(__clang__) && defined(__GNUC__) &&                 \
+     (__GNUC__ > 5 || (__GNUC__ == 5 && __GNUC_MINOR__ >= 1)) && \
+     (defined(_LIBCPP_VERSION) || defined(__GLIBCXX__))) ||      \
+    (defined(_MSC_VER) && !defined(__NVCC__))
+    #define PHMAP_HAVE_STD_IS_TRIVIALLY_CONSTRUCTIBLE 1
+    #define PHMAP_HAVE_STD_IS_TRIVIALLY_ASSIGNABLE 1
+#endif
+
+// -------------------------------------------------------------------
+// Checks whether C++11's `thread_local` storage duration specifier is
+// supported.
+// -------------------------------------------------------------------
+#ifdef PHMAP_HAVE_THREAD_LOCAL
+    #error PHMAP_HAVE_THREAD_LOCAL cannot be directly set
+#elif defined(__APPLE__)
+    #if __has_feature(cxx_thread_local) && \
+        !(TARGET_OS_IPHONE && __IPHONE_OS_VERSION_MIN_REQUIRED < __IPHONE_9_0)
+        #define PHMAP_HAVE_THREAD_LOCAL 1
+    #endif
+#else  // !defined(__APPLE__)
+    #define PHMAP_HAVE_THREAD_LOCAL 1
+#endif
+
+#if defined(__ANDROID__) && defined(__clang__)
+
+    #if __has_include(<android/ndk-version.h>)
+        #include <android/ndk-version.h>
+    #endif  // __has_include(<android/ndk-version.h>)
+
+    #if defined(__ANDROID__) && defined(__clang__) && defined(__NDK_MAJOR__) && \
+        defined(__NDK_MINOR__) &&                                               \
+        ((__NDK_MAJOR__ < 12) || ((__NDK_MAJOR__ == 12) && (__NDK_MINOR__ < 1)))
+        #undef PHMAP_HAVE_TLS
+        #undef PHMAP_HAVE_THREAD_LOCAL
+    #endif
+#endif 
+
+// ------------------------------------------------------------
+// Checks whether the __int128 compiler extension for a 128-bit 
+// integral type is supported.
+// ------------------------------------------------------------
+#ifdef PHMAP_HAVE_INTRINSIC_INT128
+    #error PHMAP_HAVE_INTRINSIC_INT128 cannot be directly set
+#elif defined(__SIZEOF_INT128__)
+    #if (defined(__clang__) && !defined(_WIN32) && !defined(__aarch64__)) || \
+        (defined(__CUDACC__) && __CUDACC_VER_MAJOR__ >= 9) ||                \
+        (defined(__GNUC__) && !defined(__clang__) && !defined(__CUDACC__))
+        #define PHMAP_HAVE_INTRINSIC_INT128 1
+    #elif defined(__CUDACC__)
+        #if __CUDACC_VER__ >= 70000
+            #define PHMAP_HAVE_INTRINSIC_INT128 1
+        #endif  // __CUDACC_VER__ >= 70000
+    #endif  // defined(__CUDACC__)
+#endif
+
+// ------------------------------------------------------------------
+// Checks whether the compiler both supports and enables exceptions. 
+// ------------------------------------------------------------------
+#ifdef PHMAP_HAVE_EXCEPTIONS
+    #error PHMAP_HAVE_EXCEPTIONS cannot be directly set.
+#elif defined(__clang__)
+    #if defined(__EXCEPTIONS) && __has_feature(cxx_exceptions)
+        #define PHMAP_HAVE_EXCEPTIONS 1
+    #endif  // defined(__EXCEPTIONS) && __has_feature(cxx_exceptions)
+#elif !(defined(__GNUC__) && (__GNUC__ < 5) && !defined(__EXCEPTIONS)) &&    \
+    !(defined(__GNUC__) && (__GNUC__ >= 5) && !defined(__cpp_exceptions)) && \
+    !(defined(_MSC_VER) && !defined(_CPPUNWIND))
+    #define PHMAP_HAVE_EXCEPTIONS 1
+#endif
+
+
+// -----------------------------------------------------------------------
+// Checks whether the platform has an mmap(2) implementation as defined in
+// POSIX.1-2001.
+// -----------------------------------------------------------------------
+#ifdef PHMAP_HAVE_MMAP
+    #error PHMAP_HAVE_MMAP cannot be directly set
+#elif defined(__linux__) || defined(__APPLE__) || defined(__FreeBSD__) ||   \
+    defined(__ros__) || defined(__native_client__) || defined(__asmjs__) || \
+    defined(__wasm__) || defined(__Fuchsia__) || defined(__sun) || \
+    defined(__ASYLO__)
+    #define PHMAP_HAVE_MMAP 1
+#endif
+
+// -----------------------------------------------------------------------
+// Checks the endianness of the platform.
+// -----------------------------------------------------------------------
+#if defined(PHMAP_IS_BIG_ENDIAN)
+    #error "PHMAP_IS_BIG_ENDIAN cannot be directly set."
+#endif
+
+#if defined(PHMAP_IS_LITTLE_ENDIAN)
+    #error "PHMAP_IS_LITTLE_ENDIAN cannot be directly set."
+#endif
+
+#if (defined(__BYTE_ORDER__) && defined(__ORDER_LITTLE_ENDIAN__) && \
+     __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)
+    #define PHMAP_IS_LITTLE_ENDIAN 1
+#elif defined(__BYTE_ORDER__) && defined(__ORDER_BIG_ENDIAN__) && \
+    __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+    #define PHMAP_IS_BIG_ENDIAN 1
+#elif defined(_WIN32)
+    #define PHMAP_IS_LITTLE_ENDIAN 1
+#else
+    #error "phmap endian detection needs to be set up for your compiler"
+#endif
+
+#if defined(__APPLE__) && defined(_LIBCPP_VERSION) && \
+    defined(__MAC_OS_X_VERSION_MIN_REQUIRED__) &&     \
+    __ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ < 101400
+    #define PHMAP_INTERNAL_MACOS_CXX17_TYPES_UNAVAILABLE 1
+#else
+    #define PHMAP_INTERNAL_MACOS_CXX17_TYPES_UNAVAILABLE 0
+#endif
+
+// ---------------------------------------------------------------------------
+// Checks whether C++17 std::any is available by checking whether <any> exists.
+// ---------------------------------------------------------------------------
+#ifdef PHMAP_HAVE_STD_ANY
+    #error "PHMAP_HAVE_STD_ANY cannot be directly set."
+#endif
+
+#ifdef __has_include
+    #if __has_include(<any>) && __cplusplus >= 201703L && \
+        !PHMAP_INTERNAL_MACOS_CXX17_TYPES_UNAVAILABLE
+        #define PHMAP_HAVE_STD_ANY 1
+    #endif
+#endif
+
+#ifdef PHMAP_HAVE_STD_OPTIONAL
+    #error "PHMAP_HAVE_STD_OPTIONAL cannot be directly set."
+#endif
+
+#ifdef __has_include
+    #if __has_include(<optional>) && __cplusplus >= 201703L && \
+        !PHMAP_INTERNAL_MACOS_CXX17_TYPES_UNAVAILABLE
+        #define PHMAP_HAVE_STD_OPTIONAL 1
+    #endif
+#endif
+
+#ifdef PHMAP_HAVE_STD_VARIANT
+    #error "PHMAP_HAVE_STD_VARIANT cannot be directly set."
+#endif
+
+#ifdef __has_include
+    #if __has_include(<variant>) && __cplusplus >= 201703L && \
+        !PHMAP_INTERNAL_MACOS_CXX17_TYPES_UNAVAILABLE
+        #define PHMAP_HAVE_STD_VARIANT 1
+    #endif
+#endif
+
+#ifdef PHMAP_HAVE_STD_STRING_VIEW
+    #error "PHMAP_HAVE_STD_STRING_VIEW cannot be directly set."
+#endif
+
+#ifdef __has_include
+    #if __has_include(<string_view>) && __cplusplus >= 201703L && \
+        (!defined(_MSC_VER) || _MSC_VER >= 1920) // vs2019
+        #define PHMAP_HAVE_STD_STRING_VIEW 1
+    #endif
+#endif
+
+// #pragma message(PHMAP_VAR_NAME_VALUE(_MSVC_LANG))
+
+#if defined(_MSC_VER) && _MSC_VER >= 1910 && PHMAP_HAVE_CC17
+    // #define PHMAP_HAVE_STD_ANY 1
+    #define PHMAP_HAVE_STD_OPTIONAL 1
+    #define PHMAP_HAVE_STD_VARIANT 1
+    #if !defined(PHMAP_HAVE_STD_STRING_VIEW) && _MSC_VER >= 1920
+        #define PHMAP_HAVE_STD_STRING_VIEW 1
+    #endif
+#endif
+
+#if PHMAP_HAVE_CC17
+    #define PHMAP_HAVE_SHARED_MUTEX 1
+#endif
+
+#ifndef PHMAP_HAVE_STD_STRING_VIEW
+    #define PHMAP_HAVE_STD_STRING_VIEW 0
+#endif
+
+// In debug mode, MSVC 2017's std::variant throws a EXCEPTION_ACCESS_VIOLATION
+// SEH exception from emplace for variant<SomeStruct> when constructing the
+// struct can throw. This defeats some of variant_test and
+// variant_exception_safety_test.
+#if defined(_MSC_VER) && _MSC_VER >= 1700 && defined(_DEBUG)
+    #define PHMAP_INTERNAL_MSVC_2017_DBG_MODE
+#endif
+
+
+// -----------------------------------------------------------------------------
+// Sanitizer Attributes
+// -----------------------------------------------------------------------------
+//
+// Sanitizer-related attributes are not "defined" in this file (and indeed
+// are not defined as such in any file). To utilize the following
+// sanitizer-related attributes within your builds, define the following macros
+// within your build using a `-D` flag, along with the given value for
+// `-fsanitize`:
+//
+//   * `ADDRESS_SANITIZER` + `-fsanitize=address` (Clang, GCC 4.8)
+//   * `MEMORY_SANITIZER` + `-fsanitize=memory` (Clang-only)
+//   * `THREAD_SANITIZER + `-fsanitize=thread` (Clang, GCC 4.8+)
+//   * `UNDEFINED_BEHAVIOR_SANITIZER` + `-fsanitize=undefined` (Clang, GCC 4.9+)
+//   * `CONTROL_FLOW_INTEGRITY` + -fsanitize=cfi (Clang-only)
+// -----------------------------------------------------------------------------
+
+// -----------------------------------------------------------------------------
+// A function-like feature checking macro that is a wrapper around
+// `__has_attribute`, which is defined by GCC 5+ and Clang and evaluates to a
+// nonzero constant integer if the attribute is supported or 0 if not.
+//
+// It evaluates to zero if `__has_attribute` is not defined by the compiler.
+// -----------------------------------------------------------------------------
+#ifdef __has_attribute
+    #define PHMAP_HAVE_ATTRIBUTE(x) __has_attribute(x)
+#else
+    #define PHMAP_HAVE_ATTRIBUTE(x) 0
+#endif
+
+// -----------------------------------------------------------------------------
+// A function-like feature checking macro that accepts C++11 style attributes.
+// It's a wrapper around `__has_cpp_attribute`, defined by ISO C++ SD-6
+// (https://en.cppreference.com/w/cpp/experimental/feature_test). If we don't
+// find `__has_cpp_attribute`, will evaluate to 0.
+// -----------------------------------------------------------------------------
+#if defined(__cplusplus) && defined(__has_cpp_attribute)
+    #define PHMAP_HAVE_CPP_ATTRIBUTE(x) __has_cpp_attribute(x)
+#else
+    #define PHMAP_HAVE_CPP_ATTRIBUTE(x) 0
+#endif
+
+// -----------------------------------------------------------------------------
+// Function Attributes
+// -----------------------------------------------------------------------------
+#if PHMAP_HAVE_ATTRIBUTE(format) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_PRINTF_ATTRIBUTE(string_index, first_to_check) \
+      __attribute__((__format__(__printf__, string_index, first_to_check)))
+    #define PHMAP_SCANF_ATTRIBUTE(string_index, first_to_check) \
+      __attribute__((__format__(__scanf__, string_index, first_to_check)))
+#else
+    #define PHMAP_PRINTF_ATTRIBUTE(string_index, first_to_check)
+    #define PHMAP_SCANF_ATTRIBUTE(string_index, first_to_check)
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(always_inline) || \
+    (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_ALWAYS_INLINE __attribute__((always_inline))
+    #define PHMAP_HAVE_ATTRIBUTE_ALWAYS_INLINE 1
+#else
+    #define PHMAP_ATTRIBUTE_ALWAYS_INLINE
+#endif
+
+#if !defined(__INTEL_COMPILER) && (PHMAP_HAVE_ATTRIBUTE(noinline) || (defined(__GNUC__) && !defined(__clang__)))
+    #define PHMAP_ATTRIBUTE_NOINLINE __attribute__((noinline))
+    #define PHMAP_HAVE_ATTRIBUTE_NOINLINE 1
+#else
+    #define PHMAP_ATTRIBUTE_NOINLINE
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(disable_tail_calls)
+    #define PHMAP_HAVE_ATTRIBUTE_NO_TAIL_CALL 1
+    #define PHMAP_ATTRIBUTE_NO_TAIL_CALL __attribute__((disable_tail_calls))
+#elif defined(__GNUC__) && !defined(__clang__)
+    #define PHMAP_HAVE_ATTRIBUTE_NO_TAIL_CALL 1
+    #define PHMAP_ATTRIBUTE_NO_TAIL_CALL \
+      __attribute__((optimize("no-optimize-sibling-calls")))
+#else
+    #define PHMAP_ATTRIBUTE_NO_TAIL_CALL
+    #define PHMAP_HAVE_ATTRIBUTE_NO_TAIL_CALL 0
+#endif
+
+#if (PHMAP_HAVE_ATTRIBUTE(weak) || \
+     (defined(__GNUC__) && !defined(__clang__))) && \
+    !(defined(__llvm__) && defined(_WIN32))
+    #undef PHMAP_ATTRIBUTE_WEAK
+    #define PHMAP_ATTRIBUTE_WEAK __attribute__((weak))
+    #define PHMAP_HAVE_ATTRIBUTE_WEAK 1
+#else
+    #define PHMAP_ATTRIBUTE_WEAK
+    #define PHMAP_HAVE_ATTRIBUTE_WEAK 0
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(nonnull) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_NONNULL(arg_index) __attribute__((nonnull(arg_index)))
+#else
+    #define PHMAP_ATTRIBUTE_NONNULL(...)
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(noreturn) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_NORETURN __attribute__((noreturn))
+#elif defined(_MSC_VER)
+    #define PHMAP_ATTRIBUTE_NORETURN __declspec(noreturn)
+#else
+    #define PHMAP_ATTRIBUTE_NORETURN
+#endif
+
+#if defined(__GNUC__) && defined(ADDRESS_SANITIZER)
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_ADDRESS __attribute__((no_sanitize_address))
+#else
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_ADDRESS
+#endif
+
+#if defined(__GNUC__) && defined(MEMORY_SANITIZER)
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_MEMORY __attribute__((no_sanitize_memory))
+#else
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_MEMORY
+#endif
+
+#if defined(__GNUC__) && defined(THREAD_SANITIZER)
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_THREAD __attribute__((no_sanitize_thread))
+#else
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_THREAD
+#endif
+
+#if defined(__GNUC__) && \
+    (defined(UNDEFINED_BEHAVIOR_SANITIZER) || defined(ADDRESS_SANITIZER))
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_UNDEFINED \
+      __attribute__((no_sanitize("undefined")))
+#else
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_UNDEFINED
+#endif
+
+#if defined(__GNUC__) && defined(CONTROL_FLOW_INTEGRITY)
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_CFI __attribute__((no_sanitize("cfi")))
+#else
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_CFI
+#endif
+
+#if defined(__GNUC__) && defined(SAFESTACK_SANITIZER)
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_SAFESTACK \
+      __attribute__((no_sanitize("safe-stack")))
+#else
+    #define PHMAP_ATTRIBUTE_NO_SANITIZE_SAFESTACK
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(returns_nonnull) || \
+    (defined(__GNUC__) && \
+     (__GNUC__ > 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 9)) && \
+     !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_RETURNS_NONNULL __attribute__((returns_nonnull))
+#else
+    #define PHMAP_ATTRIBUTE_RETURNS_NONNULL
+#endif
+
+#ifdef PHMAP_HAVE_ATTRIBUTE_SECTION
+    #error PHMAP_HAVE_ATTRIBUTE_SECTION cannot be directly set
+#elif (PHMAP_HAVE_ATTRIBUTE(section) ||                \
+       (defined(__GNUC__) && !defined(__clang__))) && \
+    !defined(__APPLE__) && PHMAP_HAVE_ATTRIBUTE_WEAK
+    #define PHMAP_HAVE_ATTRIBUTE_SECTION 1
+    #ifndef PHMAP_ATTRIBUTE_SECTION
+        #define PHMAP_ATTRIBUTE_SECTION(name) \
+          __attribute__((section(#name))) __attribute__((noinline))
+    #endif
+    #ifndef PHMAP_ATTRIBUTE_SECTION_VARIABLE
+        #define PHMAP_ATTRIBUTE_SECTION_VARIABLE(name) __attribute__((section(#name)))
+    #endif
+    #ifndef PHMAP_DECLARE_ATTRIBUTE_SECTION_VARS
+        #define PHMAP_DECLARE_ATTRIBUTE_SECTION_VARS(name) \
+          extern char __start_##name[] PHMAP_ATTRIBUTE_WEAK;    \
+          extern char __stop_##name[] PHMAP_ATTRIBUTE_WEAK
+    #endif
+    #ifndef PHMAP_DEFINE_ATTRIBUTE_SECTION_VARS
+        #define PHMAP_INIT_ATTRIBUTE_SECTION_VARS(name)
+        #define PHMAP_DEFINE_ATTRIBUTE_SECTION_VARS(name)
+    #endif
+    #define PHMAP_ATTRIBUTE_SECTION_START(name) \
+      (reinterpret_cast<void *>(__start_##name))
+    #define PHMAP_ATTRIBUTE_SECTION_STOP(name) \
+      (reinterpret_cast<void *>(__stop_##name))
+#else  // !PHMAP_HAVE_ATTRIBUTE_SECTION
+    #define PHMAP_HAVE_ATTRIBUTE_SECTION 0
+    #define PHMAP_ATTRIBUTE_SECTION(name)
+    #define PHMAP_ATTRIBUTE_SECTION_VARIABLE(name)
+    #define PHMAP_INIT_ATTRIBUTE_SECTION_VARS(name)
+    #define PHMAP_DEFINE_ATTRIBUTE_SECTION_VARS(name)
+    #define PHMAP_DECLARE_ATTRIBUTE_SECTION_VARS(name)
+    #define PHMAP_ATTRIBUTE_SECTION_START(name) (reinterpret_cast<void *>(0))
+    #define PHMAP_ATTRIBUTE_SECTION_STOP(name) (reinterpret_cast<void *>(0))
+#endif  // PHMAP_ATTRIBUTE_SECTION
+
+#if PHMAP_HAVE_ATTRIBUTE(force_align_arg_pointer) || \
+    (defined(__GNUC__) && !defined(__clang__))
+    #if defined(__i386__)
+        #define PHMAP_ATTRIBUTE_STACK_ALIGN_FOR_OLD_LIBC \
+          __attribute__((force_align_arg_pointer))
+        #define PHMAP_REQUIRE_STACK_ALIGN_TRAMPOLINE (0)
+    #elif defined(__x86_64__)
+        #define PHMAP_REQUIRE_STACK_ALIGN_TRAMPOLINE (1)
+        #define PHMAP_ATTRIBUTE_STACK_ALIGN_FOR_OLD_LIBC
+    #else  // !__i386__ && !__x86_64
+        #define PHMAP_REQUIRE_STACK_ALIGN_TRAMPOLINE (0)
+        #define PHMAP_ATTRIBUTE_STACK_ALIGN_FOR_OLD_LIBC
+    #endif  // __i386__
+#else
+    #define PHMAP_ATTRIBUTE_STACK_ALIGN_FOR_OLD_LIBC
+    #define PHMAP_REQUIRE_STACK_ALIGN_TRAMPOLINE (0)
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(nodiscard)
+    #define PHMAP_MUST_USE_RESULT [[nodiscard]]
+#elif defined(__clang__) && PHMAP_HAVE_ATTRIBUTE(warn_unused_result)
+    #define PHMAP_MUST_USE_RESULT __attribute__((warn_unused_result))
+#else
+    #define PHMAP_MUST_USE_RESULT
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(hot) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_HOT __attribute__((hot))
+#else
+    #define PHMAP_ATTRIBUTE_HOT
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(cold) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_COLD __attribute__((cold))
+#else
+    #define PHMAP_ATTRIBUTE_COLD
+#endif
+
+#if defined(__clang__)
+    #if PHMAP_HAVE_CPP_ATTRIBUTE(clang::reinitializes)
+        #define PHMAP_ATTRIBUTE_REINITIALIZES [[clang::reinitializes]]
+    #else
+        #define PHMAP_ATTRIBUTE_REINITIALIZES
+    #endif
+#else
+    #define PHMAP_ATTRIBUTE_REINITIALIZES
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(unused) || (defined(__GNUC__) && !defined(__clang__))
+    #undef PHMAP_ATTRIBUTE_UNUSED
+    #define PHMAP_ATTRIBUTE_UNUSED __attribute__((__unused__))
+#else
+    #define PHMAP_ATTRIBUTE_UNUSED
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(tls_model) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_INITIAL_EXEC __attribute__((tls_model("initial-exec")))
+#else
+    #define PHMAP_ATTRIBUTE_INITIAL_EXEC
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(packed) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_PACKED __attribute__((__packed__))
+#else
+    #define PHMAP_ATTRIBUTE_PACKED
+#endif
+
+#if PHMAP_HAVE_ATTRIBUTE(aligned) || (defined(__GNUC__) && !defined(__clang__))
+    #define PHMAP_ATTRIBUTE_FUNC_ALIGN(bytes) __attribute__((aligned(bytes)))
+#else
+    #define PHMAP_ATTRIBUTE_FUNC_ALIGN(bytes)
+#endif
+
+// ----------------------------------------------------------------------
+// Figure out SSE support
+// ----------------------------------------------------------------------
+#ifndef PHMAP_HAVE_SSE2
+    #if defined(__SSE2__) ||  \
+        (defined(_MSC_VER) && \
+         (defined(_M_X64) || (defined(_M_IX86) && _M_IX86_FP >= 2)))
+        #define PHMAP_HAVE_SSE2 1
+    #else
+        #define PHMAP_HAVE_SSE2 0
+    #endif
+#endif
+
+#ifndef PHMAP_HAVE_SSSE3
+    #if defined(__SSSE3__) || defined(__AVX2__)
+        #define PHMAP_HAVE_SSSE3 1
+    #else
+        #define PHMAP_HAVE_SSSE3 0
+    #endif
+#endif
+
+#if PHMAP_HAVE_SSSE3 && !PHMAP_HAVE_SSE2
+    #error "Bad configuration!"
+#endif
+
+#if PHMAP_HAVE_SSE2
+    #include <emmintrin.h>
+#endif
+
+#if PHMAP_HAVE_SSSE3
+    #include <tmmintrin.h>
+#endif
+
+
+// ----------------------------------------------------------------------
+// constexpr if
+// ----------------------------------------------------------------------
+#if PHMAP_HAVE_CC17
+    #define PHMAP_IF_CONSTEXPR(expr) if constexpr ((expr))
+#else 
+    #define PHMAP_IF_CONSTEXPR(expr) if ((expr))
+#endif
+
+// ----------------------------------------------------------------------
+// base/macros.h
+// ----------------------------------------------------------------------
+
+// PHMAP_ARRAYSIZE()
+//
+// Returns the number of elements in an array as a compile-time constant, which
+// can be used in defining new arrays. If you use this macro on a pointer by
+// mistake, you will get a compile-time error.
+#define PHMAP_ARRAYSIZE(array) \
+  (sizeof(::phmap::macros_internal::ArraySizeHelper(array)))
+
+namespace phmap {
+namespace macros_internal {
+    // Note: this internal template function declaration is used by PHMAP_ARRAYSIZE.
+    // The function doesn't need a definition, as we only use its type.
+    template <typename T, size_t N>
+    auto ArraySizeHelper(const T (&array)[N]) -> char (&)[N];
+}  // namespace macros_internal
+}  // namespace phmap
+
+// TODO(zhangxy): Use c++17 standard [[fallthrough]] macro, when supported.
+#if defined(__clang__) && defined(__has_warning)
+    #if __has_feature(cxx_attributes) && __has_warning("-Wimplicit-fallthrough")
+        #define PHMAP_FALLTHROUGH_INTENDED [[clang::fallthrough]]
+    #endif
+#elif defined(__GNUC__) && __GNUC__ >= 7
+    #define PHMAP_FALLTHROUGH_INTENDED [[gnu::fallthrough]]
+#endif
+
+#ifndef PHMAP_FALLTHROUGH_INTENDED
+    #define PHMAP_FALLTHROUGH_INTENDED \
+      do {  } while (0)
+#endif
+
+// PHMAP_DEPRECATED()
+//
+// Marks a deprecated class, struct, enum, function, method and variable
+// declarations. The macro argument is used as a custom diagnostic message (e.g.
+// suggestion of a better alternative).
+//
+// Example:
+//
+//   class PHMAP_DEPRECATED("Use Bar instead") Foo {...};
+//   PHMAP_DEPRECATED("Use Baz instead") void Bar() {...}
+//
+// Every usage of a deprecated entity will trigger a warning when compiled with
+// clang's `-Wdeprecated-declarations` option. This option is turned off by
+// default, but the warnings will be reported by clang-tidy.
+#if defined(__clang__) && __cplusplus >= 201103L
+    #define PHMAP_DEPRECATED(message) __attribute__((deprecated(message)))
+#endif
+
+#ifndef PHMAP_DEPRECATED
+    #define PHMAP_DEPRECATED(message)
+#endif
+
+// PHMAP_BAD_CALL_IF()
+//
+// Used on a function overload to trap bad calls: any call that matches the
+// overload will cause a compile-time error. This macro uses a clang-specific
+// "enable_if" attribute, as described at
+// http://clang.llvm.org/docs/AttributeReference.html#enable-if
+//
+// Overloads which use this macro should be bracketed by
+// `#ifdef PHMAP_BAD_CALL_IF`.
+//
+// Example:
+//
+//   int isdigit(int c);
+//   #ifdef PHMAP_BAD_CALL_IF
+//   int isdigit(int c)
+//     PHMAP_BAD_CALL_IF(c <= -1 || c > 255,
+//                       "'c' must have the value of an unsigned char or EOF");
+//   #endif // PHMAP_BAD_CALL_IF
+
+#if defined(__clang__)
+    #if __has_attribute(enable_if)
+        #define PHMAP_BAD_CALL_IF(expr, msg) \
+            __attribute__((enable_if(expr, "Bad call trap"), unavailable(msg)))
+    #endif
+#endif
+
+// PHMAP_ASSERT()
+//
+// In C++11, `assert` can't be used portably within constexpr functions.
+// PHMAP_ASSERT functions as a runtime assert but works in C++11 constexpr
+// functions.  Example:
+//
+// constexpr double Divide(double a, double b) {
+//   return PHMAP_ASSERT(b != 0), a / b;
+// }
+//
+// This macro is inspired by
+// https://akrzemi1.wordpress.com/2017/05/18/asserts-in-constexpr-functions/
+#if defined(NDEBUG)
+    #define PHMAP_ASSERT(expr) (false ? (void)(expr) : (void)0)
+#else
+    #define PHMAP_ASSERT(expr)              \
+      (PHMAP_PREDICT_TRUE((expr)) ? (void)0 \
+                                 : [] { assert(false && #expr); }())  // NOLINT
+#endif
+
+#ifdef PHMAP_HAVE_EXCEPTIONS
+    #define PHMAP_INTERNAL_TRY try
+    #define PHMAP_INTERNAL_CATCH_ANY catch (...)
+    #define PHMAP_INTERNAL_RETHROW do { throw; } while (false)
+#else  // PHMAP_HAVE_EXCEPTIONS
+    #define PHMAP_INTERNAL_TRY if (true)
+    #define PHMAP_INTERNAL_CATCH_ANY else if (false)
+    #define PHMAP_INTERNAL_RETHROW do {} while (false)
+#endif  // PHMAP_HAVE_EXCEPTIONS
+
+
+#endif // phmap_config_h_guard_

starrygl/sample/sample_core/parallel_hashmap/phmap_dump.h

+#if !defined(phmap_dump_h_guard_)
+#define phmap_dump_h_guard_
+
+// ---------------------------------------------------------------------------
+// Copyright (c) 2019, Gregory Popovitch - greg7mdp@gmail.com
+//
+//       providing dump/load/mmap_load
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+// ---------------------------------------------------------------------------
+
+#include <iostream>
+#include <fstream>
+#include <sstream>
+#include "phmap.h"
+namespace phmap
+{
+
+namespace type_traits_internal {
+
+#if defined(__GLIBCXX__) && __GLIBCXX__ < 20150801
+    template<typename T> struct IsTriviallyCopyable : public std::integral_constant<bool, __has_trivial_copy(T)> {};
+#else
+    template<typename T> struct IsTriviallyCopyable : public std::is_trivially_copyable<T> {};
+#endif
+
+template <class T1, class T2>
+struct IsTriviallyCopyable<std::pair<T1, T2>> {
+    static constexpr bool value = IsTriviallyCopyable<T1>::value && IsTriviallyCopyable<T2>::value;
+};
+}
+
+namespace priv {
+
+// ------------------------------------------------------------------------
+// dump/load for raw_hash_set
+// ------------------------------------------------------------------------
+template <class Policy, class Hash, class Eq, class Alloc>
+template<typename OutputArchive>
+bool raw_hash_set<Policy, Hash, Eq, Alloc>::dump(OutputArchive& ar) const {
+    static_assert(type_traits_internal::IsTriviallyCopyable<value_type>::value,
+                    "value_type should be trivially copyable");
+
+    if (!ar.dump(size_)) {
+        std::cerr << "Failed to dump size_" << std::endl;
+        return false;
+    }
+    if (size_ == 0) {
+        return true;
+    }
+    if (!ar.dump(capacity_)) {
+        std::cerr << "Failed to dump capacity_" << std::endl;
+        return false;
+    }
+    if (!ar.dump(reinterpret_cast<char*>(ctrl_),
+        sizeof(ctrl_t) * (capacity_ + Group::kWidth + 1))) {
+
+        std::cerr << "Failed to dump ctrl_" << std::endl;
+        return false;
+    }
+    if (!ar.dump(reinterpret_cast<char*>(slots_),
+                    sizeof(slot_type) * capacity_)) {
+        std::cerr << "Failed to dump slot_" << std::endl;
+        return false;
+    }
+    return true;
+}
+
+template <class Policy, class Hash, class Eq, class Alloc>
+template<typename InputArchive>
+bool raw_hash_set<Policy, Hash, Eq, Alloc>::load(InputArchive& ar) {
+    static_assert(type_traits_internal::IsTriviallyCopyable<value_type>::value,
+                    "value_type should be trivially copyable");
+    raw_hash_set<Policy, Hash, Eq, Alloc>().swap(*this); // clear any existing content
+    if (!ar.load(&size_)) {
+        std::cerr << "Failed to load size_" << std::endl;
+        return false;
+    }
+    if (size_ == 0) {
+        return true;
+    }
+    if (!ar.load(&capacity_)) {
+        std::cerr << "Failed to load capacity_" << std::endl;
+        return false;
+    }
+
+    // allocate memory for ctrl_ and slots_
+    initialize_slots();
+    if (!ar.load(reinterpret_cast<char*>(ctrl_),
+        sizeof(ctrl_t) * (capacity_ + Group::kWidth + 1))) {
+        std::cerr << "Failed to load ctrl" << std::endl;
+        return false;
+    }
+    if (!ar.load(reinterpret_cast<char*>(slots_),
+                    sizeof(slot_type) * capacity_)) {
+        std::cerr << "Failed to load slot" << std::endl;
+        return false;
+    }
+    return true;
+}
+
+// ------------------------------------------------------------------------
+// dump/load for parallel_hash_set
+// ------------------------------------------------------------------------
+template <size_t N,
+          template <class, class, class, class> class RefSet,
+          class Mtx_,
+          class Policy, class Hash, class Eq, class Alloc>
+template<typename OutputArchive>
+bool parallel_hash_set<N, RefSet, Mtx_, Policy, Hash, Eq, Alloc>::dump(OutputArchive& ar) const {
+    static_assert(type_traits_internal::IsTriviallyCopyable<value_type>::value,
+                    "value_type should be trivially copyable");
+
+    if (! ar.dump(subcnt())) {
+        std::cerr << "Failed to dump meta!" << std::endl;
+        return false;
+    }
+    for (size_t i = 0; i < sets_.size(); ++i) {
+        auto& inner = sets_[i];
+        typename Lockable::UniqueLock m(const_cast<Inner&>(inner));
+        if (!inner.set_.dump(ar)) {
+            std::cerr << "Failed to dump submap " << i << std::endl;
+            return false;
+        }
+    }
+    return true;
+}
+
+template <size_t N,
+          template <class, class, class, class> class RefSet,
+          class Mtx_,
+          class Policy, class Hash, class Eq, class Alloc>
+template<typename InputArchive>
+bool parallel_hash_set<N, RefSet, Mtx_, Policy, Hash, Eq, Alloc>::load(InputArchive& ar) {
+    static_assert(type_traits_internal::IsTriviallyCopyable<value_type>::value,
+                    "value_type should be trivially copyable");
+
+    size_t submap_count = 0;
+    if (!ar.load(&submap_count)) {
+        std::cerr << "Failed to load submap count!" << std::endl;
+        return false;
+    }
+
+    if (submap_count != subcnt()) {
+        std::cerr << "submap count(" << submap_count << ") != N(" << N << ")" << std::endl;
+        return false;
+    }
+
+    for (size_t i = 0; i < submap_count; ++i) {            
+        auto& inner = sets_[i];
+        typename Lockable::UniqueLock m(const_cast<Inner&>(inner));
+        if (!inner.set_.load(ar)) {
+            std::cerr << "Failed to load submap " << i << std::endl;
+            return false;
+        }
+    }
+    return true;
+}
+} // namespace priv
+
+
+
+// ------------------------------------------------------------------------
+// BinaryArchive
+//       File is closed when archive object is destroyed
+// ------------------------------------------------------------------------
+
+// ------------------------------------------------------------------------
+// ------------------------------------------------------------------------
+class BinaryOutputArchive {
+public:
+    BinaryOutputArchive(const char *file_path) {
+        ofs_.open(file_path, std::ios_base::binary);
+    }
+
+    bool dump(const char *p, size_t sz) {
+        ofs_.write(p, sz);
+        return true;
+    }
+
+    template<typename V>
+    typename std::enable_if<type_traits_internal::IsTriviallyCopyable<V>::value, bool>::type
+    dump(const V& v) {
+        ofs_.write(reinterpret_cast<const char *>(&v), sizeof(V));
+        return true;
+    }
+
+private:
+    std::ofstream ofs_;
+};
+
+
+class BinaryInputArchive {
+public:
+    BinaryInputArchive(const char * file_path) {
+        ifs_.open(file_path, std::ios_base::binary);
+    }
+
+    bool load(char* p, size_t sz) {
+        ifs_.read(p, sz);
+        return true;
+    }
+
+    template<typename V>
+    typename std::enable_if<type_traits_internal::IsTriviallyCopyable<V>::value, bool>::type
+    load(V* v) {
+        ifs_.read(reinterpret_cast<char *>(v), sizeof(V));
+        return true;
+    }
+
+private:
+    std::ifstream ifs_;
+};
+
+} // namespace phmap
+
+#endif // phmap_dump_h_guard_

starrygl/sample/sample_core/parallel_hashmap/phmap_fwd_decl.h

+#if !defined(phmap_fwd_decl_h_guard_)
+#define phmap_fwd_decl_h_guard_
+
+// ---------------------------------------------------------------------------
+// Copyright (c) 2019, Gregory Popovitch - greg7mdp@gmail.com
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+// ---------------------------------------------------------------------------
+
+#ifdef _MSC_VER
+    #pragma warning(push)  
+    #pragma warning(disable : 4514) // unreferenced inline function has been removed
+    #pragma warning(disable : 4710) // function not inlined
+    #pragma warning(disable : 4711) // selected for automatic inline expansion
+#endif
+
+#include <memory>
+#include <utility>
+
+#if defined(PHMAP_USE_ABSL_HASH) && !defined(ABSL_HASH_HASH_H_)
+    namespace absl { template <class T> struct Hash; };
+#endif
+
+namespace phmap {
+
+#if defined(PHMAP_USE_ABSL_HASH)
+    template <class T> using Hash = ::absl::Hash<T>;
+#else
+    template <class T> struct Hash;
+#endif
+
+    template <class T> struct EqualTo;
+    template <class T> struct Less;
+    template <class T> using Allocator      = typename std::allocator<T>;
+    template<class T1, class T2> using Pair = typename std::pair<T1, T2>;
+
+    class NullMutex;
+
+    namespace priv {
+
+        // The hash of an object of type T is computed by using phmap::Hash.
+        template <class T, class E = void>
+        struct HashEq 
+        {
+            using Hash = phmap::Hash<T>;
+            using Eq   = phmap::EqualTo<T>;
+        };
+
+        template <class T>
+        using hash_default_hash = typename priv::HashEq<T>::Hash;
+
+        template <class T>
+        using hash_default_eq = typename priv::HashEq<T>::Eq;
+
+        // type alias for std::allocator so we can forward declare without including other headers
+        template <class T>  
+        using Allocator = typename phmap::Allocator<T>;
+
+        // type alias for std::pair so we can forward declare without including other headers
+        template<class T1, class T2> 
+        using Pair = typename phmap::Pair<T1, T2>;
+
+    }  // namespace priv
+
+    // ------------- forward declarations for hash containers ----------------------------------
+    template <class T, 
+              class Hash  = phmap::priv::hash_default_hash<T>,
+              class Eq    = phmap::priv::hash_default_eq<T>,
+              class Alloc = phmap::priv::Allocator<T>>  // alias for std::allocator
+        class flat_hash_set;
+
+    template <class K, class V,
+              class Hash  = phmap::priv::hash_default_hash<K>,
+              class Eq    = phmap::priv::hash_default_eq<K>,
+              class Alloc = phmap::priv::Allocator<
+                            phmap::priv::Pair<const K, V>>> // alias for std::allocator
+        class flat_hash_map;
+    
+    template <class T, 
+              class Hash  = phmap::priv::hash_default_hash<T>,
+              class Eq    = phmap::priv::hash_default_eq<T>,
+              class Alloc = phmap::priv::Allocator<T>> // alias for std::allocator
+        class node_hash_set;
+
+    template <class Key, class Value,
+              class Hash  = phmap::priv::hash_default_hash<Key>,
+              class Eq    = phmap::priv::hash_default_eq<Key>,
+              class Alloc = phmap::priv::Allocator<
+                            phmap::priv::Pair<const Key, Value>>> // alias for std::allocator
+        class node_hash_map;
+
+    template <class T,
+              class Hash  = phmap::priv::hash_default_hash<T>,
+              class Eq    = phmap::priv::hash_default_eq<T>,
+              class Alloc = phmap::priv::Allocator<T>, // alias for std::allocator
+              size_t N    = 4,                  // 2**N submaps
+              class Mutex = phmap::NullMutex>   // use std::mutex to enable internal locks
+        class parallel_flat_hash_set;
+
+    template <class K, class V,
+              class Hash  = phmap::priv::hash_default_hash<K>,
+              class Eq    = phmap::priv::hash_default_eq<K>,
+              class Alloc = phmap::priv::Allocator<
+                            phmap::priv::Pair<const K, V>>, // alias for std::allocator
+              size_t N    = 4,                  // 2**N submaps
+              class Mutex = phmap::NullMutex>   // use std::mutex to enable internal locks
+        class parallel_flat_hash_map;
+
+    template <class T, 
+              class Hash  = phmap::priv::hash_default_hash<T>,
+              class Eq    = phmap::priv::hash_default_eq<T>,
+              class Alloc = phmap::priv::Allocator<T>, // alias for std::allocator
+              size_t N    = 4,                  // 2**N submaps
+              class Mutex = phmap::NullMutex>   // use std::mutex to enable internal locks
+        class parallel_node_hash_set;
+
+    template <class Key, class Value,
+              class Hash  = phmap::priv::hash_default_hash<Key>,
+              class Eq    = phmap::priv::hash_default_eq<Key>,
+              class Alloc = phmap::priv::Allocator<
+                            phmap::priv::Pair<const Key, Value>>, // alias for std::allocator
+              size_t N    = 4,                  // 2**N submaps
+              class Mutex = phmap::NullMutex>   // use std::mutex to enable internal locks
+        class parallel_node_hash_map;
+
+    // ------------- forward declarations for btree containers ----------------------------------
+    template <typename Key, typename Compare = phmap::Less<Key>,
+              typename Alloc = phmap::Allocator<Key>>
+        class btree_set;
+
+    template <typename Key, typename Compare = phmap::Less<Key>,
+              typename Alloc = phmap::Allocator<Key>>
+        class btree_multiset;
+
+    template <typename Key, typename Value, typename Compare = phmap::Less<Key>,
+              typename Alloc = phmap::Allocator<phmap::priv::Pair<const Key, Value>>>
+        class btree_map;
+    
+    template <typename Key, typename Value, typename Compare = phmap::Less<Key>,
+              typename Alloc = phmap::Allocator<phmap::priv::Pair<const Key, Value>>>
+        class btree_multimap;
+
+}  // namespace phmap
+
+
+#ifdef _MSC_VER
+     #pragma warning(pop)  
+#endif
+
+#endif // phmap_fwd_decl_h_guard_

starrygl/sample/sample_core/parallel_hashmap/phmap_utils.h

+#if !defined(phmap_utils_h_guard_)
+#define phmap_utils_h_guard_
+
+// ---------------------------------------------------------------------------
+// Copyright (c) 2019, Gregory Popovitch - greg7mdp@gmail.com
+//
+//       minimal header providing phmap::HashState
+//
+//       use as:  phmap::HashState().combine(0, _first_name, _last_name, _age);
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      https://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+// ---------------------------------------------------------------------------
+
+#ifdef _MSC_VER
+    #pragma warning(push)  
+    #pragma warning(disable : 4514) // unreferenced inline function has been removed
+    #pragma warning(disable : 4710) // function not inlined
+    #pragma warning(disable : 4711) // selected for automatic inline expansion
+#endif
+
+#include <cstdint>
+#include <functional>
+#include <tuple>
+#include "phmap_bits.h"
+
+// ---------------------------------------------------------------
+// Absl forward declaration requires global scope.
+// ---------------------------------------------------------------
+#if defined(PHMAP_USE_ABSL_HASH) && !defined(phmap_fwd_decl_h_guard_) && !defined(ABSL_HASH_HASH_H_)
+    namespace absl { template <class T> struct Hash; };
+#endif
+
+namespace phmap
+{
+
+// ---------------------------------------------------------------
+// ---------------------------------------------------------------
+template<int n> 
+struct phmap_mix
+{
+    inline size_t operator()(size_t) const;
+};
+
+template<>
+struct phmap_mix<4>
+{
+    inline size_t operator()(size_t a) const
+    {
+        static constexpr uint64_t kmul = 0xcc9e2d51UL;
+        // static constexpr uint64_t kmul = 0x3B9ACB93UL; // [greg] my own random prime
+        uint64_t l = a * kmul;
+        return static_cast<size_t>(l ^ (l >> 32));
+    }
+};
+
+#if defined(PHMAP_HAS_UMUL128)
+    template<>
+    struct phmap_mix<8>
+    {
+        // Very fast mixing (similar to Abseil)
+        inline size_t operator()(size_t a) const
+        {
+            static constexpr uint64_t k = 0xde5fb9d2630458e9ULL;
+            // static constexpr uint64_t k = 0x7C9D0BF0567102A5ULL; // [greg] my own random prime
+            uint64_t h;
+            uint64_t l = umul128(a, k, &h);
+            return static_cast<size_t>(h + l);
+        }
+    };
+#else
+    template<>
+    struct phmap_mix<8>
+    {
+        inline size_t operator()(size_t a) const
+        {
+            a = (~a) + (a << 21); // a = (a << 21) - a - 1;
+            a = a ^ (a >> 24);
+            a = (a + (a << 3)) + (a << 8); // a * 265
+            a = a ^ (a >> 14);
+            a = (a + (a << 2)) + (a << 4); // a * 21
+            a = a ^ (a >> 28);
+            a = a + (a << 31);
+            return static_cast<size_t>(a);
+        }
+    };
+#endif
+
+// --------------------------------------------
+template<int n> 
+struct fold_if_needed
+{
+    inline size_t operator()(uint64_t) const;
+};
+
+template<>
+struct fold_if_needed<4>
+{
+    inline size_t operator()(uint64_t a) const
+    {
+        return static_cast<size_t>(a ^ (a >> 32));
+    }
+};
+
+template<>
+struct fold_if_needed<8>
+{
+    inline size_t operator()(uint64_t a) const
+    {
+        return static_cast<size_t>(a);
+    }
+};
+
+// ---------------------------------------------------------------
+// see if class T has a hash_value() friend method
+// ---------------------------------------------------------------
+template<typename T>
+struct has_hash_value
+{
+private:
+    typedef std::true_type yes;
+    typedef std::false_type no;
+
+    template<typename U> static auto test(int) -> decltype(hash_value(std::declval<const U&>()) == 1, yes());
+
+    template<typename> static no test(...);
+
+public:
+    static constexpr bool value = std::is_same<decltype(test<T>(0)), yes>::value;
+};
+
+#if defined(PHMAP_USE_ABSL_HASH) && !defined(phmap_fwd_decl_h_guard_)
+    template <class T> using Hash = ::absl::Hash<T>;
+#elif !defined(PHMAP_USE_ABSL_HASH)
+// ---------------------------------------------------------------
+//               phmap::Hash
+// ---------------------------------------------------------------
+template <class T>
+struct Hash
+{
+    template <class U, typename std::enable_if<has_hash_value<U>::value, int>::type = 0>
+    size_t _hash(const T& val) const
+    {
+        return hash_value(val);
+    }
+ 
+    template <class U, typename std::enable_if<!has_hash_value<U>::value, int>::type = 0>
+    size_t _hash(const T& val) const
+    {
+        return std::hash<T>()(val);
+    }
+ 
+    inline size_t operator()(const T& val) const
+    {
+        return _hash<T>(val);
+    }
+};
+
+template <class T>
+struct Hash<T *>
+{
+    inline size_t operator()(const T *val) const noexcept
+    {
+        return static_cast<size_t>(reinterpret_cast<const uintptr_t>(val)); 
+    }
+};
+
+template<class ArgumentType, class ResultType>
+struct phmap_unary_function
+{
+    typedef ArgumentType argument_type;
+    typedef ResultType result_type;
+};
+
+template <>
+struct Hash<bool> : public phmap_unary_function<bool, size_t>
+{
+    inline size_t operator()(bool val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<char> : public phmap_unary_function<char, size_t>
+{
+    inline size_t operator()(char val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<signed char> : public phmap_unary_function<signed char, size_t>
+{
+    inline size_t operator()(signed char val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<unsigned char> : public phmap_unary_function<unsigned char, size_t>
+{
+    inline size_t operator()(unsigned char val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<wchar_t> : public phmap_unary_function<wchar_t, size_t>
+{
+    inline size_t operator()(wchar_t val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<int16_t> : public phmap_unary_function<int16_t, size_t>
+{
+    inline size_t operator()(int16_t val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<uint16_t> : public phmap_unary_function<uint16_t, size_t>
+{
+    inline size_t operator()(uint16_t val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<int32_t> : public phmap_unary_function<int32_t, size_t>
+{
+    inline size_t operator()(int32_t val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<uint32_t> : public phmap_unary_function<uint32_t, size_t>
+{
+    inline size_t operator()(uint32_t val) const noexcept
+    { return static_cast<size_t>(val); }
+};
+
+template <>
+struct Hash<int64_t> : public phmap_unary_function<int64_t, size_t>
+{
+    inline size_t operator()(int64_t val) const noexcept
+    { return fold_if_needed<sizeof(size_t)>()(static_cast<uint64_t>(val)); }
+};
+
+template <>
+struct Hash<uint64_t> : public phmap_unary_function<uint64_t, size_t>
+{
+    inline size_t operator()(uint64_t val) const noexcept
+    { return fold_if_needed<sizeof(size_t)>()(val); }
+};
+
+template <>
+struct Hash<float> : public phmap_unary_function<float, size_t>
+{
+    inline size_t operator()(float val) const noexcept
+    {
+        // -0.0 and 0.0 should return same hash
+        uint32_t *as_int = reinterpret_cast<uint32_t *>(&val);
+        return (val == 0) ? static_cast<size_t>(0) : 
+                            static_cast<size_t>(*as_int);
+    }
+};
+
+template <>
+struct Hash<double> : public phmap_unary_function<double, size_t>
+{
+    inline size_t operator()(double val) const noexcept
+    {
+        // -0.0 and 0.0 should return same hash
+        uint64_t *as_int = reinterpret_cast<uint64_t *>(&val);
+        return (val == 0) ? static_cast<size_t>(0) : 
+                            fold_if_needed<sizeof(size_t)>()(*as_int);
+    }
+};
+
+#endif
+
+template <class H, int sz> struct Combiner
+{
+    H operator()(H seed, size_t value);
+};
+
+template <class H> struct Combiner<H, 4>
+{
+    H operator()(H seed, size_t value)
+    {
+        return seed ^ (value + 0x9e3779b9 + (seed << 6) + (seed >> 2));
+    }
+};
+
+template <class H> struct Combiner<H, 8>
+{
+    H operator()(H seed, size_t value)
+    {
+        return seed ^ (value + size_t(0xc6a4a7935bd1e995) + (seed << 6) + (seed >> 2));
+    }
+};
+
+// define HashState to combine member hashes... see example below
+// -----------------------------------------------------------------------------
+template <typename H>
+class HashStateBase {
+public:
+    template <typename T, typename... Ts>
+    static H combine(H state, const T& value, const Ts&... values);
+
+    static H combine(H state) { return state; }
+};
+
+template <typename H>
+template <typename T, typename... Ts>
+H HashStateBase<H>::combine(H seed, const T& v, const Ts&... vs)
+{
+    return HashStateBase<H>::combine(Combiner<H, sizeof(H)>()(
+                                         seed, phmap::Hash<T>()(v)), 
+                                     vs...);
+}
+
+using HashState = HashStateBase<size_t>;
+
+// -----------------------------------------------------------------------------
+
+#if !defined(PHMAP_USE_ABSL_HASH)
+
+// define Hash for std::pair
+// -------------------------
+template<class T1, class T2> 
+struct Hash<std::pair<T1, T2>> {
+    size_t operator()(std::pair<T1, T2> const& p) const noexcept {
+        return phmap::HashState().combine(phmap::Hash<T1>()(p.first), p.second);
+    }
+};
+
+// define Hash for std::tuple
+// --------------------------
+template<class... T> 
+struct Hash<std::tuple<T...>> {
+    size_t operator()(std::tuple<T...> const& t) const noexcept {
+        return _hash_helper(t);
+    }
+
+private:
+    template<size_t I = 0, class ...P>
+        typename std::enable_if<I == sizeof...(P), size_t>::type
+        _hash_helper(const std::tuple<P...> &) const noexcept { return 0; }
+
+    template<size_t I = 0, class ...P>
+    typename std::enable_if<I < sizeof...(P), size_t>::type
+    _hash_helper(const std::tuple<P...> &t) const noexcept {
+        const auto &el = std::get<I>(t);
+        using el_type = typename std::remove_cv<typename std::remove_reference<decltype(el)>::type>::type;
+        return Combiner<size_t, sizeof(size_t)>()(
+            phmap::Hash<el_type>()(el),  _hash_helper<I + 1>(t));
+    }
+};
+
+
+#endif
+
+
+}  // namespace phmap
+
+#ifdef _MSC_VER
+     #pragma warning(pop)  
+#endif
+
+#endif // phmap_utils_h_guard_

starrygl/sample/sample_core/random_walk_sampler.py

+import torch
+import torch.multiprocessing as mp
+from typing import Optional, Tuple
+
+from base import BaseSampler, NegativeSampling, SampleOutput
+from neighbor_sampler import NeighborSampler, SampleType
+
+class RandomWalkSampler(BaseSampler):
+    def __init__(
+        self,
+        num_nodes: int,
+        num_layers: int,
+        graph_data,
+        workers = 1,
+        tnb = None,
+        is_distinct = 0,
+        policy = "uniform",
+        edge_weight: Optional[torch.Tensor] = None,
+        graph_name = None
+    ) -> None:
+        r"""__init__
+        Args:
+            num_nodes: the num of all nodes in the graph
+            num_layers: the num of layers to be sampled
+            workers: the number of threads, default value is 1
+            tnb: neighbor infomation table
+            is_distinct: 1-need distinct, 0-don't need distinct
+            policy: "uniform" or "recent" or "weighted"
+            is_root_ts: 1-base on root's ts, 0-base on parent node's ts
+            edge_weight: the initial weights of edges
+            graph_name: the name of graph
+        """
+        super().__init__()        
+        self.sampler = NeighborSampler(
+                        num_nodes=num_nodes, 
+                        tnb=tnb,
+                        num_layers=num_layers, 
+                        fanout=[1 for i in range(num_layers)], 
+                        graph_data=graph_data,
+                        edge_weight = edge_weight, 
+                        workers=workers, 
+                        policy=policy,
+                        graph_name=graph_name,
+                        is_distinct = is_distinct
+                        )
+        self.num_layers = num_layers
+
+    def _get_sample_info(self):
+        return self.num_nodes,self.num_layers,self.fanout,self.workers
+    
+    def _get_sample_options(self):
+        return {"is_distinct" : self.is_distinct,
+                "policy" : self.policy,
+                "with_eid" : self.tnb.with_eid, 
+                "weighted" : self.tnb.weighted, 
+                "with_timestamp" : self.tnb.with_timestamp}
+    
+    def insert_edges_with_timestamp(
+            self, 
+            edge_index : torch.Tensor, 
+            eid : torch.Tensor, 
+            timestamp : torch.Tensor,
+            edge_weight : Optional[torch.Tensor] = None):
+        row, col = edge_index
+        # 更新节点数和tnb
+        self.num_nodes = self.tnb.update_neighbors_with_time(
+            row.contiguous(), 
+            col.contiguous(), 
+            timestamp.contiguous(), 
+            eid.contiguous(), 
+            self.is_distinct, 
+            edge_weight.contiguous())
+    
+    def update_edges_weight(
+            self, 
+            edge_index : torch.Tensor, 
+            eid : torch.Tensor,
+            edge_weight : Optional[torch.Tensor] = None):
+        row, col = edge_index
+        # 更新tnb的权重信息
+        if self.tnb.with_eid:
+            self.tnb.update_edge_weight(
+                eid.contiguous(),
+                col.contiguous(),
+                edge_weight.contiguous()
+            )
+        else:
+            self.tnb.update_edge_weight(
+                row.contiguous(),
+                col.contiguous(),
+                edge_weight.contiguous()
+            )
+    
+    def update_nodes_weight(
+            self, 
+            nid : torch.Tensor,
+            node_weight : Optional[torch.Tensor] = None):
+        # 更新tnb的权重信息
+        self.tnb.update_node_weight(
+            nid.contiguous(),
+            node_weight.contiguous()
+        )
+
+    def update_all_node_weight(
+            self,
+            node_weight : torch.Tensor):
+        # 更新tnb的权重信息
+        self.tnb.update_all_node_weight(node_weight.contiguous())
+    
+    def sample_from_nodes(
+        self,
+        nodes: torch.Tensor,
+        with_outer_sample: SampleType,
+        ts: Optional[torch.Tensor] = None
+    ) -> SampleOutput:
+        r"""Performs mutilayer sampling from the nodes specified in: nodes
+        The specific number of layers is determined by parameter: num_layers
+        returning a sampled subgraph in the specified output format: Tuple[torch.Tensor, list].
+
+        Args:
+            nodes: the list of seed nodes index
+            with_outer_sample: 0-sample in whole graph structure; 1-sample onehop outer nodel; 2-cross partition sampling
+        Returns:
+            sampled_nodes: the node sampled
+            sampled_edge_index: the edge sampled
+        """
+        return self.sampler.sample_from_nodes(nodes, ts, with_outer_sample)
+    
+    def sample_from_edges(
+        self,
+        edges: torch.Tensor,
+        ets: Optional[torch.Tensor] = None,
+        neg_sampling: Optional[NegativeSampling] = None, 
+        with_outer_sample: SampleType = SampleType.Whole
+    ) -> SampleOutput:
+        r"""Performs sampling from the edges specified in :obj:`index`,
+        returning a sampled subgraph in the specified output format.
+
+        Args:
+            edges: the list of seed edges index
+            with_outer_sample: 0-sample in whole graph structure; 1-sample onehop outer nodel; 2-cross partition sampling
+            edge_label: the label for the seed edges.
+            neg_sampling: The negative sampling configuration
+        Returns:
+            sampled_nodes: the nodes sampled
+            sampled_edge_index_list: the edges sampled
+        """
+        return self.sampler.sample_from_edges(edges, ets, neg_sampling, with_outer_sample)
+
+
+if __name__=="__main__":
+    edge_index1 = torch.tensor([[0, 1, 1, 1, 1, 2, 2, 2, 2, 4, 4, 4, 5], # , 3, 3
+                                [1, 0, 2, 0, 4, 1, 3, 0, 3, 3, 5, 0, 2]])# , 2, 5
+    timeStamp=torch.FloatTensor([1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4])
+    edge_weight1 = None
+    num_nodes1 = 6
+    num_neighbors = 2
+    # Run the neighbor sampling
+    from Utils import GraphData
+    g_data = GraphData(id=0, edge_index=edge_index1, timestamp=timeStamp, data=None, partptr=torch.tensor([0, num_nodes1]))
+    # Run the random walk sampling
+    sampler=RandomWalkSampler(num_nodes=num_nodes1, 
+                              num_layers=3,
+                              edge_weight=edge_weight1, 
+                              graph_data=g_data, 
+                              graph_name='a',
+                              workers=4,
+                              is_root_ts=0,
+                              is_distinct = 0)
+
+    out = sampler.sample_from_nodes(torch.tensor([1,2]),
+                                    with_outer_sample=SampleType.Whole, 
+                                    ts=torch.tensor([1, 2]))
+    # out = sampler.sample_from_edges(torch.tensor([[1,2],[4,0]]), 
+    #                                 with_outer_sample=SampleType.Whole, 
+    #                                 ets = torch.tensor([1, 2]))
+    
+    # Print the result
+    print('node:', out.node)
+    print('edge_index_list:', out.edge_index_list)
+    print('eid_list:', out.eid_list)
+    print('eid_ts_list:', out.eid_ts_list)
+    print('metadata: ', out.metadata)
+

starrygl/sample/sample_core/reddit_demo_before.py

+import argparse
+import random
+import pandas as pd
+import numpy as np
+import torch
+from ogb.nodeproppred import PygNodePropPredDataset
+from torch_geometric.datasets import Reddit
+import time
+
+from tqdm import tqdm
+from Utils import GraphData
+
+class NegLinkSampler:
+
+    def __init__(self, num_nodes):
+        self.num_nodes = num_nodes
+
+    def sample(self, n):
+        return np.random.randint(self.num_nodes, size=n)
+
+class NegLinkInductiveSampler:
+    def __init__(self, nodes):
+        self.nodes = list(nodes)
+
+    def sample(self, n):
+        return np.random.choice(self.nodes, size=n)
+
+def load_reddit_dataset(data_path):
+    df = pd.read_csv('/home/zlj/hzq/project/code/TGL/DATA/{}/edges.csv'.format("REDDIT"))
+    num_nodes = max(int(df['src'].max()), int(df['dst'].max())) + 1
+    src = torch.tensor(df['src'].to_numpy(dtype=int))
+    dst = torch.tensor(df['dst'].to_numpy(dtype=int))
+    edge_index = torch.stack([src, dst])
+    timestamp = torch.tensor(df['time']).float()
+    g = GraphData(0, edge_index, timestamp=timestamp, data=None, partptr=torch.tensor([0, num_nodes]))
+    return g, df
+
+
+
+parser=argparse.ArgumentParser()
+parser.add_argument('--data', type=str, help='dataset name',default="REDDIT")
+parser.add_argument('--config', type=str, help='path to config file',default="/home/zlj/hzq/project/code/TGL/config/TGN.yml")
+parser.add_argument('--batch_size', type=int, default=600, help='path to config file')
+parser.add_argument('--num_thread', type=int, default=64, help='number of thread')
+args=parser.parse_args()
+
+seed=10
+torch.manual_seed(seed) # 为CPU设置随机种子
+torch.cuda.manual_seed(seed) # 为当前GPU设置随机种子
+torch.cuda.manual_seed_all(seed)  # if you are using multi-GPU，为所有GPU设置随机种子
+np.random.seed(seed)  # Numpy module.
+random.seed(seed)  # Python random module.	
+torch.backends.cudnn.benchmark = False
+torch.backends.cudnn.deterministic = True
+
+g_data, df = load_reddit_dataset("/home/zlj/hzq/code/dataset/reddit")
+print(g_data)
+# for worker in [1,2,3,4,5,6,7,8,9,10,20,30]:
+# import random
+# timestamp = [random.randint(1, 5) for i in range(0, g.num_edges)]
+# timestamp = torch.FloatTensor(timestamp)
+
+# print('begin load')
+# pre = time.time()
+# # timestamp = torch.load('/home/zlj/hzq/code/gnn/my_sampler/TemporalSample/timestamp.my')
+# tnb = torch.load("tnb_reddit_before.my")
+# end = time.time()
+# print("load time:", end-pre)
+# row, col = g.edge_index
+edge_weight=None
+# g_data = GraphData(id=1, edge_index=g.edge_index, timestamp=timestamp, data=g, partptr=torch.tensor([0, g.num_nodes//4, g.num_nodes//4*2, g.num_nodes//4*3, g.num_nodes]))
+
+from neighbor_sampler import NeighborSampler, SampleType, get_neighbors
+
+print('begin tnb')
+row, col = g_data.edge_index
+row = torch.cat([row, col])
+col = torch.cat([col, row])
+eid = torch.cat([g_data.eid, g_data.eid])
+timestamp = torch.cat([g_data.edge_ts, g_data.edge_ts])
+timestamp,ind = timestamp.sort()
+timestamp = timestamp.float().contiguous()
+eid  = eid[ind].contiguous()
+row = row[ind]
+col = col[ind]
+print(row, col)
+pre = time.time()
+tnb = get_neighbors("reddit", row.contiguous(), col.contiguous(), g_data.num_nodes, 0, eid, edge_weight, timestamp)
+end = time.time()
+print("init tnb time:", end-pre)
+torch.save(tnb, "tnb_reddit_before.my")
+
+
+pre = time.time()
+sampler = NeighborSampler(g_data.num_nodes, 
+                          tnb=tnb,
+                          num_layers=1, 
+                          fanout=[10], 
+                          graph_data=g_data, 
+                          workers=10, 
+                          policy="recent", 
+                          graph_name='a')
+end = time.time()
+print("init time:", end-pre)
+
+# neg_link_sampler = NegLinkSampler(g_data.num_nodes)
+from base import NegativeSampling, NegativeSamplingMode
+neg_link_sampler = NegativeSampling(NegativeSamplingMode.triplet)
+
+# from torch_geometric.sampler import NeighborSampler, NumNeighbors, NodeSamplerInput, SamplerOutput
+# pre = time.time()
+# num_nei = NumNeighbors([100, 100])
+# node_idx = NodeSamplerInput(input_id=None, node=torch.tensor(range(g.num_nodes//4, g.num_nodes//4+600000)))# (input_id=None, node=torch.masked_select(torch.arange(g.num_nodes),node_data['train_mask']))
+# sampler = NeighborSampler(g, num_nei)
+# end = time.time()
+# print("init time:", end-pre)
+
+
+out = []
+tot_time = 0
+sam_time = 0
+sam_edge = 0
+pre = time.time()
+
+
+for _, rows in tqdm(df.groupby(df.index // args.batch_size), total=len(df) // args.batch_size):
+        # root_nodes = torch.tensor(np.concatenate([rows.src.values, rows.dst.values, neg_link_sampler.sample(len(rows))])).long()
+        # ts = torch.tensor(np.concatenate([rows.time.values, rows.time.values, rows.time.values]).astype(np.float32))
+        # outi = sampler.sample_from_nodes(root_nodes, ts=ts)
+        edges = torch.tensor(np.stack([rows.src.values, rows.dst.values])).long()
+        outi, meta = sampler.sample_from_edges(edges=edges, ets=torch.tensor(rows.time.values).float(), neg_sampling=neg_link_sampler)
+        tot_time += outi[0].tot_time
+        sam_time += outi[0].sample_time
+        # print(outi[0].sample_edge_num)
+        sam_edge += outi[0].sample_edge_num
+        out.append(outi)
+
+end = time.time()
+print("sample time", end-pre)
+print("tot_time", tot_time)
+print("sam_time", sam_time)
+print("sam_edge", sam_edge)
+print('eid_list:', out[23][0].eid())
+# print('delta_ts_list:', out[10][0].delta_ts)
+print('node:', out[23][0].sample_nodes())
+# print('node_ts:', out[23][0].sample_nodes_ts)
+# print('eid_list:', out[23][1].eid)
+# print('node:', out[23][1].sample_nodes)
+# print('node_ts:', out[23][1].sample_nodes_ts)
+# print('edge_index_list:', out[0][0].edge_index)
\ No newline at end of file

starrygl/sample/sample_core/sample_cores_dist.cpp

+#include <iostream>
+#include <omp.h>
+#include <pybind11/pybind11.h>
+#include <pybind11/numpy.h>
+#include <pybind11/stl.h>
+#include <torch/extension.h>
+#include <time.h>
+#include <random>
+#include <phmap.h>
+#include <boost/thread/mutex.hpp>
+#define MTX boost::mutex
+
+using namespace std;
+namespace py = pybind11;
+namespace th = torch;
+
+typedef int64_t NodeIDType;
+typedef int64_t EdgeIDType;
+typedef float WeightType;
+typedef float TimeStampType;
+
+
+#define EXTRAARGS , phmap::priv::hash_default_hash<K>, \
+                            phmap::priv::hash_default_eq<K>, \
+                            std::allocator<K>, 4, MTX
+template <class K>
+using HashT  = phmap::parallel_flat_hash_set<K EXTRAARGS>;
+
+template <class K, class V>
+using HashM  = phmap::parallel_flat_hash_map<K, V EXTRAARGS>;
+
+
+class TemporalNeighborBlock;
+class TemporalGraphBlock;
+TemporalNeighborBlock& get_neighbors(string graph_name, th::Tensor row, th::Tensor col, int64_t num_nodes, int is_distinct, optional<th::Tensor> eid, optional<th::Tensor> edge_weight, optional<th::Tensor> time);
+th::Tensor heads_unique(th::Tensor array, th::Tensor heads, int threads);
+int nodeIdToInOut(NodeIDType nid, int pid, const vector<NodeIDType>& part_ptr);
+int nodeIdToPartId(NodeIDType nid, const vector<NodeIDType>& part_ptr);
+vector<th::Tensor> divide_nodes_to_part(th::Tensor nodes, const vector<NodeIDType>& part_ptr, int threads);
+// vector<int64_t> sample_multinomial(vector<WeightType> weights, int num_samples, bool replacement, default_random_engine e);
+NodeIDType sample_multinomial(const vector<WeightType>& weights, default_random_engine& e);
+
+
+template<typename T>
+inline py::array vec2npy(const std::vector<T> &vec)
+{
+    // need to let python garbage collector handle C++ vector memory 
+    // see https://github.com/pybind/pybind11/issues/1042
+    // non-copy value transfer
+    auto v = new std::vector<T>(vec);
+    auto capsule = py::capsule(v, [](void *v)
+                               { delete reinterpret_cast<std::vector<T> *>(v); });
+    return py::array(v->size(), v->data(), capsule);
+    // return py::array(vec.size(), vec.data());
+}
+
+/* 
+ * NeighborSampler Utils
+ */
+class TemporalNeighborBlock
+{
+    public:
+        vector<vector<NodeIDType>> neighbors;
+        vector<vector<TimeStampType>> timestamp;
+        vector<vector<EdgeIDType>> eid;
+        vector<vector<WeightType>> edge_weight;
+        vector<phmap::parallel_flat_hash_map<NodeIDType, int64_t>> inverted_index;
+        vector<int64_t> deg;
+        vector<phmap::parallel_flat_hash_set<NodeIDType>> neighbors_set;
+
+        bool with_eid = false;
+        bool weighted = false;
+        bool with_timestamp = false;
+
+        TemporalNeighborBlock(){}
+        // TemporalNeighborBlock(const TemporalNeighborBlock &tnb);
+        TemporalNeighborBlock(vector<vector<NodeIDType>>& neighbors, 
+                              vector<int64_t> &deg):
+                              neighbors(neighbors), deg(deg){}
+        TemporalNeighborBlock(vector<vector<NodeIDType>>& neighbors, 
+                              vector<vector<WeightType>>& edge_weight, 
+                              vector<vector<EdgeIDType>>& eid,
+                              vector<int64_t> &deg):
+                              neighbors(neighbors), edge_weight(edge_weight),eid(eid), deg(deg)
+                              { this->with_eid=true;
+                                this->weighted=true; }
+        TemporalNeighborBlock(vector<vector<NodeIDType>>& neighbors, 
+                              vector<vector<WeightType>>& edge_weight,
+                              vector<vector<TimeStampType>>& timestamp,
+                              vector<vector<EdgeIDType>>& eid, 
+                              vector<int64_t> &deg):
+                              neighbors(neighbors), edge_weight(edge_weight), timestamp(timestamp),eid(eid), deg(deg)
+                              { this->with_eid=true;
+                                this->weighted=true; 
+                                this->with_timestamp=true;}
+        py::array get_node_neighbor(NodeIDType node_id){
+            return vec2npy(neighbors[node_id]);
+        }
+        py::array get_node_neighbor_timestamp(NodeIDType node_id){
+            return vec2npy(timestamp[node_id]);
+        }
+        int64_t get_node_deg(NodeIDType node_id){
+            return deg[node_id];
+        }
+        bool empty(){
+            return this->deg.empty();
+        }
+        void update_edge_weight(th::Tensor row_or_eid, th::Tensor col, th::Tensor edge_weight);
+        void update_node_weight(th::Tensor nid, th::Tensor node_weight);
+        void update_all_node_weight(th::Tensor node_weight);
+        int64_t update_neighbors_with_time(th::Tensor row, th::Tensor col, th::Tensor time, th::Tensor eid, int is_distinct, std::optional<th::Tensor> edge_weight);
+
+        
+        std::string serialize() const {
+            std::ostringstream oss;
+            // 序列化基本类型成员
+            oss << with_eid << " " << weighted << " " << with_timestamp << " ";
+
+            // 序列化 vector<vector<T>> 类型成员
+            auto serializeVecVec = [&oss](const auto& vecVec) {
+                for (const auto& vec : vecVec) {
+                    oss << vec.size() << " ";
+                    for (const auto& elem : vec) {
+                        oss << elem << " ";
+                    }
+                }
+                oss << "|";  // 添加一个分隔符以区分不同的 vector
+            };
+
+            serializeVecVec(neighbors);
+            serializeVecVec(timestamp);
+            serializeVecVec(eid);
+            serializeVecVec(edge_weight);
+
+            // 序列化 vector<int64_t> 类型成员
+            oss << deg.size() << " ";
+            for (const auto& d : deg) {
+                oss << d << " ";
+            }
+            oss << "|";
+
+            // 序列化 inverted_index
+            for (const auto& map : inverted_index) {
+                oss << map.size() << " ";
+                for (const auto& [key, value] : map) {
+                    oss << key << " " << value << " ";
+                }
+            }
+            oss << "|";
+
+            // 序列化 neighbors_set
+            for (const auto& set : neighbors_set) {
+                oss << set.size() << " ";
+                for (const auto& elem : set) {
+                    oss << elem << " ";
+                }
+            }
+            oss << "|";
+
+            return oss.str();
+        }
+
+        static TemporalNeighborBlock deserialize(const std::string& s) {
+            std::istringstream iss(s);
+            TemporalNeighborBlock tnb;
+
+            // 反序列化基本类型成员
+            iss >> tnb.with_eid >> tnb.weighted >> tnb.with_timestamp;
+
+            // 反序列化 vector<vector<T>> 类型成员
+            auto deserializeVecLong = [&iss](vector<vector<int64_t>>& vecVec) {
+                std::string segment;
+                std::getline(iss, segment, '|');
+                std::istringstream vec_iss(segment);
+                while (!vec_iss.eof()) {
+                    size_t vec_size;
+                    vec_iss >> vec_size;
+                    if (vec_iss.eof()) break;  // 防止多余的空白
+                    vector<int64_t> vec(vec_size);
+                    for (size_t i = 0; i < vec_size; ++i) {
+                        vec_iss >> vec[i];
+                    }
+                    vecVec.push_back(vec);
+                }
+            };
+
+            
+            auto deserializeVecFloat = [&iss](vector<vector<float>>& vecVec) {
+                std::string segment;
+                std::getline(iss, segment, '|');
+                std::istringstream vec_iss(segment);
+                while (!vec_iss.eof()) {
+                    size_t vec_size;
+                    vec_iss >> vec_size;
+                    if (vec_iss.eof()) break;  // 防止多余的空白
+                    vector<float> vec(vec_size);
+                    for (size_t i = 0; i < vec_size; ++i) {
+                        vec_iss >> vec[i];
+                    }
+                    vecVec.push_back(vec);
+                }
+            };
+
+            deserializeVecLong(tnb.neighbors);
+            deserializeVecFloat(tnb.timestamp);
+            deserializeVecLong(tnb.eid);
+            deserializeVecFloat(tnb.edge_weight);
+
+            std::string segment;
+            // 反序列化 vector<int64_t> 类型成员
+            segment="";
+            std::getline(iss, segment, '|');
+            std::istringstream vec_iss(segment);
+            size_t vec_size;
+            vec_iss >> vec_size;
+            tnb.deg.resize(vec_size);
+            for (size_t i = 0; i < vec_size; ++i) {
+                vec_iss >> tnb.deg[i];
+            }
+
+            // 反序列化 inverted_index
+            segment="";
+            std::getline(iss, segment, '|');
+            std::istringstream map_iss(segment);
+            while (!map_iss.eof()) {
+                size_t map_size;
+                map_iss >> map_size;
+                if (map_iss.eof()) break;
+                phmap::parallel_flat_hash_map<NodeIDType, int64_t> map;
+                for (size_t i = 0; i < map_size; ++i) {
+                    NodeIDType key;
+                    int64_t value;
+                    map_iss >> key >> value;
+                    map[key] = value;
+                }
+                tnb.inverted_index.push_back(map);
+            }
+
+            // 反序列化 neighbors_set
+            std::getline(iss, segment, '|');
+            std::istringstream set_iss(segment);
+            while (!set_iss.eof()) {
+                size_t set_size;
+                set_iss >> set_size;
+                if (set_iss.eof()) break;
+                phmap::parallel_flat_hash_set<NodeIDType> set;
+                for (size_t i = 0; i < set_size; ++i) {
+                    NodeIDType elem;
+                    set_iss >> elem;
+                    set.insert(elem);
+                }
+                tnb.neighbors_set.push_back(set);
+            }
+
+            return tnb;
+        }
+};
+
+class TemporalGraphBlock
+{
+    public:
+        vector<NodeIDType> row;
+        vector<NodeIDType> col;
+        vector<EdgeIDType> eid;
+        vector<TimeStampType> delta_ts;
+        vector<int64_t> src_index;
+        vector<NodeIDType> sample_nodes;
+        vector<TimeStampType> sample_nodes_ts;
+        double sample_time = 0;
+        double tot_time = 0;
+        int64_t sample_edge_num = 0;
+
+        TemporalGraphBlock(){}
+        // TemporalGraphBlock(const TemporalGraphBlock &tgb);
+        TemporalGraphBlock(vector<NodeIDType> &_row, vector<NodeIDType> &_col,
+                           vector<NodeIDType> &_sample_nodes):
+                           row(_row), col(_col), sample_nodes(_sample_nodes){}
+        TemporalGraphBlock(vector<NodeIDType> &_row, vector<NodeIDType> &_col,
+                           vector<NodeIDType> &_sample_nodes,
+                           vector<TimeStampType> &_sample_nodes_ts):
+                           row(_row), col(_col), sample_nodes(_sample_nodes),
+                           sample_nodes_ts(_sample_nodes_ts){}
+};
+
+class T_TemporalGraphBlock
+{
+    public:
+        th::Tensor row;
+        th::Tensor col;
+        th::Tensor eid;
+        th::Tensor delta_ts;
+        th::Tensor src_index;
+        th::Tensor sample_nodes;
+        th::Tensor sample_nodes_ts;
+        double sample_time = 0;
+        double tot_time = 0;
+        int64_t sample_edge_num = 0;
+
+        T_TemporalGraphBlock(){}
+        T_TemporalGraphBlock(th::Tensor &_row, th::Tensor &_col,
+                           th::Tensor &_sample_nodes):
+                           row(_row), col(_col), sample_nodes(_sample_nodes){}
+};
+
+// 辅助函数
+template <typename T>
+T* get_data_ptr(const th::Tensor& tensor) {
+    AT_ASSERTM(tensor.is_contiguous(), "Offset tensor must be contiguous");
+    AT_ASSERTM(tensor.dim() == 1, "Offset tensor must be one-dimensional");
+    return tensor.data_ptr<T>();
+}
+template <typename T>
+torch::Tensor vecToTensor(const std::vector<T>& vec) {
+    // 确定数据类型
+    torch::ScalarType dtype;
+    if (std::is_same<T, int64_t>::value) {
+        dtype = torch::kInt64;
+    } else if (std::is_same<T, float>::value) {
+        dtype = torch::kFloat32;
+    } else {
+        throw std::runtime_error("Unsupported data type");
+    }
+
+    // 创建Tensor
+    torch::Tensor tensor = torch::from_blob(
+        const_cast<T*>(vec.data()), /* 数据指针 */
+        {static_cast<long>(vec.size())}, /* 尺寸 */
+        dtype /* 数据类型 */
+    );
+
+    return tensor;//.clone(); // 克隆Tensor以拷贝数据
+}
+
+TemporalNeighborBlock& get_neighbors(
+        string graph_name, th::Tensor row, th::Tensor col, int64_t num_nodes, int is_distinct, optional<th::Tensor> eid, optional<th::Tensor> edge_weight, optional<th::Tensor> time)
+{   //row、col、time按time升序排列，由时间早的到时间晚的
+    auto src = get_data_ptr<NodeIDType>(row);
+    auto dst = get_data_ptr<NodeIDType>(col);
+    EdgeIDType* eid_ptr = eid ? get_data_ptr<EdgeIDType>(eid.value()) : nullptr;
+    WeightType* ew = edge_weight ? get_data_ptr<WeightType>(edge_weight.value()) : nullptr;
+    TimeStampType* t = time ? get_data_ptr<TimeStampType>(time.value()) : nullptr;
+    int64_t edge_num = row.size(0);
+    static phmap::parallel_flat_hash_map<string, TemporalNeighborBlock> tnb_map;
+    if(tnb_map.count(graph_name)==1)
+        return tnb_map[graph_name];
+    tnb_map[graph_name] = TemporalNeighborBlock();
+    TemporalNeighborBlock& tnb = tnb_map[graph_name];
+
+    double start_time = omp_get_wtime();
+    //初始化
+    tnb.neighbors.resize(num_nodes);
+    tnb.deg.resize(num_nodes, 0);
+        //初始化optional相关
+    tnb.with_eid = eid.has_value();
+    tnb.weighted = edge_weight.has_value();
+    tnb.with_timestamp = time.has_value();
+    if (tnb.with_eid) tnb.eid.resize(num_nodes);
+    if (tnb.weighted) {
+        tnb.edge_weight.resize(num_nodes);
+        tnb.inverted_index.resize(num_nodes);
+    }
+    if (tnb.with_timestamp) tnb.timestamp.resize(num_nodes);
+        
+    //计算, 条件判断移出循环优化执行效率
+    for(int64_t i=0; i<edge_num; i++){
+        //计算节点邻居
+        tnb.neighbors[dst[i]].emplace_back(src[i]);
+    }
+    //如果有eid，插入
+    if(tnb.with_eid)
+        for(int64_t i=0; i<edge_num; i++){
+            tnb.eid[dst[i]].emplace_back(eid_ptr[i]);
+        }
+    //如果有权重信息，插入节点与邻居边的权重和反向索引
+    if(tnb.weighted)
+        for(int64_t i=0; i<edge_num; i++){
+            tnb.edge_weight[dst[i]].emplace_back(ew[i]);
+            if(tnb.with_eid) tnb.inverted_index[dst[i]][eid_ptr[i]]=tnb.neighbors[dst[i]].size()-1;
+            else tnb.inverted_index[dst[i]][src[i]]=tnb.neighbors[dst[i]].size()-1;
+        }
+    //如果有时序信息，插入节点与邻居边的时间
+    if(tnb.with_timestamp)
+        for(int64_t i=0; i<edge_num; i++){
+            tnb.timestamp[dst[i]].emplace_back(t[i]);
+        }
+        
+    if(is_distinct){
+        for(int64_t i=0; i<num_nodes; i++){
+            //收集单边去重节点度
+            phmap::parallel_flat_hash_set<NodeIDType> temp_s;
+            temp_s.insert(tnb.neighbors[i].begin(), tnb.neighbors[i].end());
+            tnb.neighbors_set.emplace_back(temp_s);
+            tnb.deg[i] = tnb.neighbors_set[i].size();
+        }
+    }
+    else{
+        for(int64_t i=0; i<num_nodes; i++){
+            //收集单边节点度
+            tnb.deg[i] = tnb.neighbors[i].size();
+        }
+    }
+    double end_time = omp_get_wtime();
+    cout<<"get_neighbors consume: "<<end_time-start_time<<"s"<<endl;
+    return tnb;
+}
+
+void TemporalNeighborBlock::update_edge_weight(
+        th::Tensor row_or_eid, th::Tensor col, th::Tensor edge_weight){
+    AT_ASSERTM(this->weighted, "This Graph has no edge weight infomation");
+    auto dst = get_data_ptr<NodeIDType>(col);
+    WeightType* ew = get_data_ptr<WeightType>(edge_weight);
+    NodeIDType* src;
+    EdgeIDType* eid_ptr;
+    if(this->with_eid) eid_ptr = get_data_ptr<EdgeIDType>(row_or_eid);
+    else src = get_data_ptr<NodeIDType>(row_or_eid);
+    
+    int64_t edge_num = col.size(0);
+
+    for(int64_t i=0; i<edge_num; i++){
+        //修改节点与邻居边的权重
+        AT_ASSERTM(this->inverted_index[dst[i]].count(src[i])==1, "Unexist Edge Index: "+to_string(src[i])+", "+to_string(dst[i]));
+		int index;
+        if(this->with_eid) index = this->inverted_index[dst[i]][eid_ptr[i]];
+        else index = this->inverted_index[dst[i]][src[i]];
+        this->edge_weight[dst[i]][index] = ew[i];
+    }
+}
+
+void TemporalNeighborBlock:: update_node_weight(th::Tensor nid, th::Tensor node_weight){
+    AT_ASSERTM(this->weighted, "This Graph has no edge weight infomation");
+    auto dst = get_data_ptr<NodeIDType>(nid);
+    WeightType* nw = get_data_ptr<WeightType>(node_weight);
+    
+    int64_t node_num = nid.size(0);
+
+    for(int64_t i=0; i<node_num; i++){
+        //修改节点与邻居边的权重
+        AT_ASSERTM(dst[i]<this->deg.size(), "Unexist Node Index: "+to_string(dst[i]));
+        if(this->inverted_index[dst[i]].empty())
+            return;
+        for(auto index : this->inverted_index[dst[i]]){
+            this->edge_weight[dst[i]][index.second] = nw[i];
+        }
+    }
+}
+
+void TemporalNeighborBlock:: update_all_node_weight(th::Tensor node_weight){
+    AT_ASSERTM(this->weighted, "This Graph has no edge weight infomation");
+    WeightType* nw = get_data_ptr<WeightType>(node_weight);
+    int64_t node_num = node_weight.size(0);
+    AT_ASSERTM(node_num==this->neighbors.size(), "The tensor node_weight size is not suitable node number.");
+
+    for(int64_t i=0; i<node_num; i++){
+        //修改节点与邻居边的权重
+        for(int j=0; j<this->neighbors[i].size();j++){
+            this->edge_weight[i][j] = nw[this->neighbors[i][j]];
+        }
+    }
+}
+
+int64_t TemporalNeighborBlock::update_neighbors_with_time(
+        th::Tensor row, th::Tensor col, th::Tensor time,th::Tensor eid, int is_distinct, std::optional<th::Tensor> edge_weight){
+        //row、col、time按time升序排列，由时间早的到时间晚的
+    AT_ASSERTM(this->empty(), "Empty TemporalNeighborBlock, please use get_neighbors_with_time");
+    AT_ASSERTM(this->with_timestamp == true, "This Graph has no time infomation!");
+    auto src = get_data_ptr<NodeIDType>(row);
+    auto dst = get_data_ptr<NodeIDType>(col);
+    auto eid_ptr = get_data_ptr<EdgeIDType>(eid);
+    auto t = get_data_ptr<TimeStampType>(time);
+    WeightType* ew = edge_weight ? get_data_ptr<WeightType>(edge_weight.value()) : nullptr;
+    int64_t edge_num = row.size(0);
+    int64_t num_nodes = this->neighbors.size();
+
+    //处理optional的值
+    if(edge_weight.has_value()){
+        AT_ASSERTM(this->weighted == true, "This Graph has no edge weight");
+    }
+    if(this->weighted){
+        AT_ASSERTM(edge_weight.has_value(), "This Graph need edge weight");
+    }
+
+    // double start_time = omp_get_wtime();
+    if(is_distinct){
+        for(int64_t i=0; i<edge_num; i++){
+            //如果有新节点
+            if(dst[i]>=num_nodes){
+                num_nodes = dst[i]+1;
+                this->neighbors.resize(num_nodes);
+                this->deg.resize(num_nodes, 0);
+                this->eid.resize(num_nodes);
+                this->timestamp.resize(num_nodes);
+                    //初始化optional相关
+                if (this->weighted) {
+                    this->edge_weight.resize(num_nodes);
+                    this->inverted_index.resize(num_nodes);
+                }
+            }
+            //更新节点邻居
+            this->neighbors[dst[i]].emplace_back(src[i]);
+            //插入eid
+            this->eid[dst[i]].emplace_back(eid_ptr[i]);
+            //插入节点与邻居边的时间
+            this->timestamp[dst[i]].emplace_back(t[i]);
+            //如果有权重信息，插入节点与邻居边的权重和反向索引
+            if(this->weighted){
+                this->edge_weight[dst[i]].emplace_back(ew[i]);
+                if(this->with_eid) this->inverted_index[dst[i]][eid_ptr[i]]=this->neighbors[dst[i]].size()-1;
+                else this->inverted_index[dst[i]][src[i]]=this->neighbors[dst[i]].size()-1;
+            }
+            
+            this->neighbors_set[dst[i]].insert(src[i]);
+            this->deg[dst[i]]=this->neighbors_set[dst[i]].size();
+        }
+    }
+    else{
+        for(int64_t i=0; i<edge_num; i++){
+            //更新节点邻居
+            this->neighbors[dst[i]].emplace_back(src[i]);
+            //插入eid
+            this->eid[dst[i]].emplace_back(eid_ptr[i]);
+            //插入节点与邻居边的时间
+            this->timestamp[dst[i]].emplace_back(t[i]);
+            //如果有权重信息，插入节点与邻居边的权重和反向索引
+            if(this->weighted){
+                this->edge_weight[dst[i]].emplace_back(ew[i]);
+                this->inverted_index[dst[i]][src[i]]=this->neighbors[dst[i]].size()-1;
+            }
+
+            this->deg[dst[i]]=this->neighbors[dst[i]].size();
+        }
+    }
+    // double end_time = omp_get_wtime();
+    // cout<<"update_neighbors consume: "<<end_time-start_time<<"s"<<endl;
+    return num_nodes;
+}
+
+class ParallelSampler
+{
+    public:
+        TemporalNeighborBlock& tnb;
+        NodeIDType num_nodes;
+        EdgeIDType num_edges;
+        int threads;
+        vector<int> fanouts;
+        // vector<NodeIDType> part_ptr;
+        // int pid;
+        int num_layers;
+        string policy;
+        std::vector<TemporalGraphBlock> ret;
+
+        ParallelSampler(TemporalNeighborBlock& _tnb, NodeIDType _num_nodes, EdgeIDType _num_edges, int _threads, 
+                        vector<int>& _fanouts, int _num_layers, string _policy) :
+                        tnb(_tnb), num_nodes(_num_nodes), num_edges(_num_edges), threads(_threads), 
+                        fanouts(_fanouts), num_layers(_num_layers), policy(_policy)
+        {
+            omp_set_num_threads(_threads);
+            ret.clear();
+            ret.resize(_num_layers);
+        }
+
+        void reset()
+        {
+            ret.clear();
+            ret.resize(num_layers);
+        }
+
+        void neighbor_sample_from_nodes(th::Tensor nodes, optional<th::Tensor> root_ts);
+        void neighbor_sample_from_nodes_static(th::Tensor nodes);
+        void neighbor_sample_from_nodes_static_layer(th::Tensor nodes, int cur_layer);
+        void neighbor_sample_from_nodes_with_before(th::Tensor nodes, th::Tensor root_ts);
+        void neighbor_sample_from_nodes_with_before_layer(th::Tensor nodes, th::Tensor root_ts, int cur_layer);
+};
+
+
+
+void ParallelSampler :: neighbor_sample_from_nodes(th::Tensor nodes, optional<th::Tensor> root_ts)
+{
+    omp_set_num_threads(threads);
+    if(policy == "weighted")
+        AT_ASSERTM(tnb.weighted, "Tnb has no weight infomation!");
+    else if(policy == "recent")
+        AT_ASSERTM(tnb.with_timestamp, "Tnb has no timestamp infomation!");
+    else if(policy == "uniform")
+        ;
+    else{
+        throw runtime_error("The policy \"" + policy + "\" is not exit!");
+    }
+    if(tnb.with_timestamp){
+        AT_ASSERTM(tnb.with_timestamp, "Tnb has no timestamp infomation!");
+        AT_ASSERTM(root_ts.has_value(), "Parameter mismatch!");
+        neighbor_sample_from_nodes_with_before(nodes, root_ts.value());
+    }
+    else{
+        neighbor_sample_from_nodes_static(nodes);
+    }
+}
+
+void ParallelSampler :: neighbor_sample_from_nodes_static_layer(th::Tensor nodes, int cur_layer){
+    py::gil_scoped_release release;
+    double tot_start_time = omp_get_wtime();
+
+    TemporalGraphBlock tgb = TemporalGraphBlock();
+    int fanout = fanouts[cur_layer];
+    ret[cur_layer] = TemporalGraphBlock();
+    auto nodes_data = get_data_ptr<NodeIDType>(nodes);    
+    vector<phmap::parallel_flat_hash_set<NodeIDType>> node_s_threads(threads);
+    phmap::parallel_flat_hash_set<NodeIDType> node_s;
+    vector<vector<NodeIDType>> eid_threads(threads);//row_threads(threads),col_threads(threads);
+    AT_ASSERTM(tnb.with_eid, "Tnb has no eid infomation! We need eid!");
+    
+    // double start_time = omp_get_wtime();
+    int reserve_capacity = int(ceil(nodes.size(0) / threads)) * fanout;
+#pragma omp parallel
+{
+    int tid = omp_get_thread_num();
+    unsigned int loc_seed = tid;
+    // row_threads[tid].reserve(reserve_capacity);
+    // col_threads[tid].reserve(reserve_capacity);
+    eid_threads[tid].reserve(reserve_capacity);
+#pragma omp for schedule(static, int(ceil(static_cast<float>((nodes.size(0)) / threads))))
+    for(int64_t i=0; i<nodes.size(0); i++){
+        // int tid = omp_get_thread_num();
+        NodeIDType node = nodes_data[i];
+        vector<NodeIDType> nei(tnb.neighbors[node]);
+        vector<EdgeIDType> edge;
+        edge = tnb.eid[node];
+
+        double s_start_time = omp_get_wtime();
+        if(tnb.deg[node]>fanout){
+            phmap::flat_hash_set<NodeIDType> temp_s;
+            default_random_engine e(8);//(time(0));
+            uniform_int_distribution<> u(0, tnb.deg[node]-1);
+            while(temp_s.size()!=fanout){
+                //循环选择fanout个邻居
+                NodeIDType indice;
+                if(policy == "weighted"){//考虑边权重信息
+                    const vector<WeightType>& ew = tnb.edge_weight[node];
+                    indice = sample_multinomial(ew, e);
+                }
+                else if(policy == "uniform"){//均匀采样
+                    indice = u(e);
+                }
+                auto chosen_n_iter = nei.begin() + indice;
+                auto rst = temp_s.insert(*chosen_n_iter);
+                if(rst.second){ //不重复
+                    auto chosen_e_iter = edge.begin() + indice;
+                    eid_threads[tid].emplace_back(*chosen_e_iter);
+                    node_s_threads[tid].insert(*chosen_n_iter);
+                }
+            }
+            // row_threads[tid].insert(row_threads[tid].end(),temp_s.begin(),temp_s.end());
+            // col_threads[tid].insert(col_threads[tid].end(), fanout, node);
+        }
+        else{
+            node_s_threads[tid].insert(nei.begin(), nei.end());
+            // row_threads[tid].insert(row_threads[tid].end(),nei.begin(),nei.end());
+            // col_threads[tid].insert(col_threads[tid].end(), tnb.deg[node], node);
+            eid_threads[tid].insert(eid_threads[tid].end(),edge.begin(), edge.end());
+        }
+        if(tid==0)
+            ret[0].sample_time += omp_get_wtime() - s_start_time;
+    }
+}
+    // double end_time = omp_get_wtime();
+    
+    // cout<<"neighbor_sample_from_nodes parallel part consume: "<<end_time-start_time<<"s"<<endl;
+    
+    int size = 0;
+    vector<int> each_begin(threads);
+    for(int i = 0; i<threads; i++){
+        int s = eid_threads[i].size();
+        each_begin[i]=size;
+        size += s;
+    }
+    // ret[cur_layer].row.resize(size);
+    // ret[cur_layer].col.resize(size);
+    ret[cur_layer].eid.resize(size);
+#pragma omp parallel for schedule(static, 1)
+    for(int i = 0; i<threads; i++){
+        // copy(row_threads[i].begin(), row_threads[i].end(), ret[cur_layer].row.begin()+each_begin[i]);
+        // copy(col_threads[i].begin(), col_threads[i].end(), ret[cur_layer].col.begin()+each_begin[i]);
+        copy(eid_threads[i].begin(), eid_threads[i].end(), ret[cur_layer].eid.begin()+each_begin[i]);
+    }
+    for(int i = 0; i<threads; i++)
+        node_s.insert(node_s_threads[i].begin(), node_s_threads[i].end());
+    ret[cur_layer].sample_nodes.assign(node_s.begin(), node_s.end());
+
+    ret[0].tot_time += omp_get_wtime() - tot_start_time;
+    ret[0].sample_edge_num += ret[cur_layer].eid.size();
+    py::gil_scoped_acquire acquire;
+}
+
+void ParallelSampler :: neighbor_sample_from_nodes_static(th::Tensor nodes){
+    for(int i=0;i<num_layers;i++){
+        if(i==0) neighbor_sample_from_nodes_static_layer(nodes, i);
+        else neighbor_sample_from_nodes_static_layer(vecToTensor<NodeIDType>(ret[i-1].sample_nodes), i);
+    }
+}
+
+void ParallelSampler :: neighbor_sample_from_nodes_with_before_layer(
+        th::Tensor nodes, th::Tensor root_ts, int cur_layer){
+    py::gil_scoped_release release;
+    double tot_start_time = omp_get_wtime();
+    ret[cur_layer] = TemporalGraphBlock();
+    auto nodes_data = get_data_ptr<NodeIDType>(nodes);
+    auto ts_data = get_data_ptr<TimeStampType>(root_ts);
+    int fanout = fanouts[cur_layer];
+    // HashT<pair<NodeIDType,TimeStampType> > node_s;
+    vector<TemporalGraphBlock> tgb_i(threads);
+    
+    default_random_engine e(8);//(time(0));
+    // double start_time = omp_get_wtime();
+    int reserve_capacity = int(ceil(nodes.size(0) / threads)) * fanout;
+#pragma omp parallel
+{
+    int tid = omp_get_thread_num();
+    unsigned int loc_seed = tid;
+    tgb_i[tid].sample_nodes.reserve(reserve_capacity);
+    tgb_i[tid].sample_nodes_ts.reserve(reserve_capacity);
+    tgb_i[tid].delta_ts.reserve(reserve_capacity);
+    tgb_i[tid].eid.reserve(reserve_capacity);
+    tgb_i[tid].src_index.reserve(reserve_capacity);
+#pragma omp for schedule(static, int(ceil(static_cast<float>((nodes.size(0)) / threads))))
+    for(int64_t i=0; i<nodes.size(0); i++){
+        // int tid = omp_get_thread_num();
+        NodeIDType node = nodes_data[i];
+        TimeStampType rtts = ts_data[i];
+        int end_index = lower_bound(tnb.timestamp[node].begin(), tnb.timestamp[node].end(), rtts)-tnb.timestamp[node].begin();
+        // cout<<node<<" "<<end_index<<" "<<tnb.deg[node]<<endl;
+
+        double s_start_time = omp_get_wtime();
+        if ((policy == "recent") || (end_index <= fanout)){
+            int start_index = max(0, end_index-fanout);
+            tgb_i[tid].src_index.insert(tgb_i[tid].src_index.end(), end_index-start_index, i);
+            tgb_i[tid].sample_nodes.insert(tgb_i[tid].sample_nodes.end(), tnb.neighbors[node].begin()+start_index, tnb.neighbors[node].begin()+end_index);
+            tgb_i[tid].sample_nodes_ts.insert(tgb_i[tid].sample_nodes_ts.end(), tnb.timestamp[node].begin()+start_index, tnb.timestamp[node].begin()+end_index);
+            tgb_i[tid].eid.insert(tgb_i[tid].eid.end(), tnb.eid[node].begin()+start_index, tnb.eid[node].begin()+end_index);
+            for(int cid = start_index; cid < end_index;cid++){
+                tgb_i[tid].delta_ts.emplace_back(rtts-tnb.timestamp[node][cid]);
+            }
+        }
+        else{
+            //可选邻居边大于扇出的话需要随机选择fanout个邻居
+            tgb_i[tid].src_index.insert(tgb_i[tid].src_index.end(), fanout, i);
+            uniform_int_distribution<> u(0, end_index-1);
+            //cout<<end_index<<endl;
+            // cout<<"start:"<<start_index<<" end:"<<end_index<<endl;
+            for(int i=0; i<fanout;i++){
+                int cid;
+                if(policy == "uniform")
+                    cid = u(e);
+                    // cid = rand_r(&loc_seed) % (end_index);
+                else if(policy == "weighted"){
+                    const vector<WeightType>& ew = tnb.edge_weight[node];
+                    cid = sample_multinomial(ew, e);
+                }
+                tgb_i[tid].sample_nodes.emplace_back(tnb.neighbors[node][cid]);
+                tgb_i[tid].sample_nodes_ts.emplace_back(tnb.timestamp[node][cid]);
+                tgb_i[tid].delta_ts.emplace_back(rtts-tnb.timestamp[node][cid]);
+                tgb_i[tid].eid.emplace_back(tnb.eid[node][cid]);
+            }
+        }
+        if(tid==0)
+            ret[0].sample_time += omp_get_wtime() - s_start_time;
+    }
+}
+    // double end_time = omp_get_wtime();
+    // cout<<"neighbor_sample_from_nodes parallel part consume: "<<end_time-start_time<<"s"<<endl;
+    
+    // start_time = omp_get_wtime();
+
+    int size = 0;
+    vector<int> each_begin(threads);
+    for(int i = 0; i<threads; i++){
+        int s = tgb_i[i].eid.size();
+        each_begin[i]=size;
+        size += s;
+    }
+    ret[cur_layer].eid.resize(size);
+    ret[cur_layer].src_index.resize(size);
+    ret[cur_layer].delta_ts.resize(size);
+    ret[cur_layer].sample_nodes.resize(size);
+    ret[cur_layer].sample_nodes_ts.resize(size);
+#pragma omp parallel for schedule(static, 1)
+    for(int i = 0; i<threads; i++){
+        copy(tgb_i[i].eid.begin(), tgb_i[i].eid.end(), ret[cur_layer].eid.begin()+each_begin[i]);
+        copy(tgb_i[i].src_index.begin(), tgb_i[i].src_index.end(), ret[cur_layer].src_index.begin()+each_begin[i]);
+        copy(tgb_i[i].delta_ts.begin(), tgb_i[i].delta_ts.end(), ret[cur_layer].delta_ts.begin()+each_begin[i]);
+        copy(tgb_i[i].sample_nodes.begin(), tgb_i[i].sample_nodes.end(), ret[cur_layer].sample_nodes.begin()+each_begin[i]);
+        copy(tgb_i[i].sample_nodes_ts.begin(), tgb_i[i].sample_nodes_ts.end(), ret[cur_layer].sample_nodes_ts.begin()+each_begin[i]);
+    }
+
+    // end_time = omp_get_wtime();
+    // cout<<"end union consume: "<<end_time-start_time<<"s"<<endl;
+
+    ret[0].tot_time += omp_get_wtime() - tot_start_time;
+    ret[0].sample_edge_num += ret[cur_layer].eid.size();
+    py::gil_scoped_acquire acquire;
+}
+
+void ParallelSampler :: neighbor_sample_from_nodes_with_before(th::Tensor nodes, th::Tensor root_ts){
+    for(int i=0;i<num_layers;i++){
+        if(i==0) neighbor_sample_from_nodes_with_before_layer(nodes, root_ts, i);
+        else neighbor_sample_from_nodes_with_before_layer(vecToTensor<NodeIDType>(ret[i-1].sample_nodes), 
+                                                          vecToTensor<TimeStampType>(ret[i-1].sample_nodes_ts), i);
+    }
+}
+
+/*-------------------------------------------------------------------------------------**
+**------------Utils--------------------------------------------------------------------**
+**-------------------------------------------------------------------------------------*/
+
+th::Tensor heads_unique(th::Tensor array, th::Tensor heads, int threads){
+    auto array_ptr = array.data_ptr<NodeIDType>();
+    phmap::parallel_flat_hash_set<NodeIDType> s(array_ptr, array_ptr+array.numel());
+    if(heads.numel()==0) return th::tensor(vector<NodeIDType>(s.begin(), s.end()));
+    AT_ASSERTM(heads.is_contiguous(), "Offset tensor must be contiguous");
+    AT_ASSERTM(heads.dim() == 1, "0ffset tensor must be one-dimensional");
+    auto heads_ptr = heads.data_ptr<NodeIDType>();
+#pragma omp parallel for num_threads(threads)
+    for(int64_t i=0; i<heads.size(0); i++){
+        if(s.count(heads_ptr[i])==1){
+        #pragma omp critical(erase)
+            s.erase(heads_ptr[i]);
+        }
+    }
+    vector<NodeIDType> ret;
+    ret.reserve(s.size()+heads.numel());
+    ret.assign(heads_ptr, heads_ptr+heads.numel());
+    ret.insert(ret.end(), s.begin(), s.end());
+    // cout<<"s: "<<s.size()<<" array: "<<array.size()<<endl;
+    return th::tensor(ret);
+}
+
+int nodeIdToPartId(NodeIDType nid, const vector<NodeIDType>& part_ptr){
+    int partitionId = -1;
+    for(int i=0;i<part_ptr.size()-1;i++){
+            if(nid>=part_ptr[i]&&nid<part_ptr[i+1]){
+                partitionId = i;
+                break;
+            }
+    }
+    if(partitionId<0) throw "nid 不存在对应的分区";
+    return partitionId;
+}
+//0:inner; 1:outer
+int nodeIdToInOut(NodeIDType nid, int pid, const vector<NodeIDType>& part_ptr){
+    if(nid>=part_ptr[pid]&&nid<part_ptr[pid+1]){
+        return 0;
+    }
+    return 1;
+}
+
+vector<th::Tensor> divide_nodes_to_part(
+        th::Tensor nodes, const vector<NodeIDType>& part_ptr, int threads){
+    double start_time = omp_get_wtime();
+    AT_ASSERTM(nodes.is_contiguous(), "Offset tensor must be contiguous");
+    AT_ASSERTM(nodes.dim() == 1, "0ffset tensor must be one-dimensional");
+    auto nodes_id = nodes.data_ptr<NodeIDType>();
+    vector<vector<vector<NodeIDType>>> node_part_threads;
+    vector<th::Tensor> result(part_ptr.size()-1);
+    //初始化点的分区，每个分区按线程划分避免冲突
+    for(int i = 0; i<threads; i++){
+        vector<vector<NodeIDType>> node_parts;
+        for(int j=0;j<part_ptr.size()-1;j++){
+            node_parts.push_back(vector<NodeIDType>());
+        }
+        node_part_threads.push_back(node_parts);
+    }
+#pragma omp parallel for num_threads(threads) default(shared)
+    for(int64_t i=0; i<nodes.size(0); i++){
+        int tid = omp_get_thread_num();
+        int pid = nodeIdToPartId(nodes_id[i], part_ptr);
+        node_part_threads[tid][pid].emplace_back(nodes_id[i]);
+    }
+#pragma omp parallel for num_threads(part_ptr.size()-1) default(shared)
+    for(int i = 0; i<part_ptr.size()-1; i++){
+        vector<NodeIDType> temp;
+        for(int j=0;j<threads;j++){
+            temp.insert(temp.end(), node_part_threads[j][i].begin(), node_part_threads[j][i].end());
+        }
+        result[i]=th::tensor(temp);
+    }
+    double end_time = omp_get_wtime();
+    // cout<<"end divide consume: "<<end_time-start_time<<"s"<<endl;
+    return result;
+}
+
+NodeIDType sample_multinomial(const vector<WeightType>& weights, default_random_engine& e){
+    NodeIDType sample_indice;
+    vector<WeightType> cumulative_weights;
+    partial_sum(weights.begin(), weights.end(), back_inserter(cumulative_weights));
+    AT_ASSERTM(cumulative_weights.back() > 0, "Edge weight sum should be greater than 0.");
+    
+    uniform_real_distribution<WeightType> distribution(0.0, cumulative_weights.back());
+    WeightType random_value = distribution(e);
+    auto it = lower_bound(cumulative_weights.begin(), cumulative_weights.end(), random_value);
+    sample_indice = distance(cumulative_weights.begin(), it);
+    return sample_indice;
+}
+
+/*------------Python Bind--------------------------------------------------------------*/
+PYBIND11_MODULE(sample_cores, m)
+{
+    m
+    .def("get_neighbors", 
+        &get_neighbors, 
+        py::return_value_policy::reference)    
+    .def("heads_unique", 
+        &heads_unique, 
+        py::return_value_policy::reference)
+    .def("divide_nodes_to_part", 
+        &divide_nodes_to_part, 
+        py::return_value_policy::reference);
+
+    py::class_<TemporalGraphBlock>(m, "TemporalGraphBlock")
+        .def(py::init<vector<NodeIDType> &, vector<NodeIDType> &,
+                      vector<NodeIDType> &>())
+        .def("row", [](const TemporalGraphBlock &tgb) { return vecToTensor<NodeIDType>(tgb.row); })
+        .def("col", [](const TemporalGraphBlock &tgb) { return vecToTensor<NodeIDType>(tgb.col); })
+        .def("eid", [](const TemporalGraphBlock &tgb) { return vecToTensor<EdgeIDType>(tgb.eid); })
+        .def("delta_ts", [](const TemporalGraphBlock &tgb) { return vecToTensor<TimeStampType>(tgb.delta_ts); })
+        .def("src_index", [](const TemporalGraphBlock &tgb) { return vecToTensor<EdgeIDType>(tgb.src_index); })
+        .def("sample_nodes", [](const TemporalGraphBlock &tgb) { return vecToTensor<NodeIDType>(tgb.sample_nodes); })
+        .def("sample_nodes_ts", [](const TemporalGraphBlock &tgb) { return vecToTensor<TimeStampType>(tgb.sample_nodes_ts); })
+        .def_readonly("sample_time", &TemporalGraphBlock::sample_time, py::return_value_policy::reference)
+        .def_readonly("tot_time", &TemporalGraphBlock::tot_time, py::return_value_policy::reference)
+        .def_readonly("sample_edge_num", &TemporalGraphBlock::sample_edge_num, py::return_value_policy::reference);
+
+    py::class_<T_TemporalGraphBlock>(m, "T_TemporalGraphBlock")
+        .def(py::init<th::Tensor &, th::Tensor &,
+                      th::Tensor &>())
+        .def_readonly("row", &T_TemporalGraphBlock::row, py::return_value_policy::reference)
+        .def_readonly("col", &T_TemporalGraphBlock::col, py::return_value_policy::reference)
+        .def_readonly("eid", &T_TemporalGraphBlock::eid, py::return_value_policy::reference)
+        .def_readonly("delta_ts", &T_TemporalGraphBlock::delta_ts, py::return_value_policy::reference)
+        .def_readonly("src_index", &T_TemporalGraphBlock::src_index, py::return_value_policy::reference)
+        .def_readonly("sample_nodes", &T_TemporalGraphBlock::sample_nodes, py::return_value_policy::reference)
+        .def_readonly("sample_nodes_ts", &T_TemporalGraphBlock::sample_nodes_ts, py::return_value_policy::reference)
+        .def_readonly("sample_time", &T_TemporalGraphBlock::sample_time, py::return_value_policy::reference)
+        .def_readonly("tot_time", &T_TemporalGraphBlock::tot_time, py::return_value_policy::reference)
+        .def_readonly("sample_edge_num", &T_TemporalGraphBlock::sample_edge_num, py::return_value_policy::reference);
+
+    py::class_<TemporalNeighborBlock>(m, "TemporalNeighborBlock")
+        .def(py::init<vector<vector<NodeIDType>>&, 
+                      vector<int64_t> &>())
+        .def(py::pickle(
+            [](const TemporalNeighborBlock& tnb) { return tnb.serialize(); },
+            [](const std::string& s) { return TemporalNeighborBlock::deserialize(s); }
+        ))
+        .def("update_neighbors_with_time", 
+            &TemporalNeighborBlock::update_neighbors_with_time)
+        .def("update_edge_weight", 
+            &TemporalNeighborBlock::update_edge_weight)
+        .def("update_node_weight", 
+            &TemporalNeighborBlock::update_node_weight)
+        .def("update_all_node_weight", 
+            &TemporalNeighborBlock::update_all_node_weight)            
+        // .def("get_node_neighbor",&TemporalNeighborBlock::get_node_neighbor)
+        // .def("get_node_deg", &TemporalNeighborBlock::get_node_deg)
+        .def_readonly("neighbors", &TemporalNeighborBlock::neighbors, py::return_value_policy::reference)
+        .def_readonly("timestamp", &TemporalNeighborBlock::timestamp, py::return_value_policy::reference)
+        .def_readonly("edge_weight", &TemporalNeighborBlock::edge_weight, py::return_value_policy::reference)
+        .def_readonly("eid", &TemporalNeighborBlock::eid, py::return_value_policy::reference)
+        .def_readonly("deg", &TemporalNeighborBlock::deg, py::return_value_policy::reference)
+        .def_readonly("with_eid", &TemporalNeighborBlock::with_eid, py::return_value_policy::reference)
+        .def_readonly("with_timestamp", &TemporalNeighborBlock::with_timestamp, py::return_value_policy::reference)
+        .def_readonly("weighted", &TemporalNeighborBlock::weighted, py::return_value_policy::reference);
+
+    py::class_<ParallelSampler>(m, "ParallelSampler")
+        .def(py::init<TemporalNeighborBlock &, NodeIDType, EdgeIDType, int,
+                      vector<int>&, int, string>())
+        .def_readonly("ret", &ParallelSampler::ret, py::return_value_policy::reference)
+        .def("neighbor_sample_from_nodes", &ParallelSampler::neighbor_sample_from_nodes)
+        .def("reset", &ParallelSampler::reset)
+        .def("get_ret", [](const ParallelSampler &ps) { return ps.ret; });
+}
\ No newline at end of file

starrygl/sample/sample_core/setup.py

+from setuptools import setup
+from torch.utils.cpp_extension import BuildExtension, CppExtension
+
+setup(
+    name='sample_cores',
+    ext_modules=[
+        CppExtension(
+            name='sample_cores',
+            sources=['sample_cores_dist.cpp'],
+            extra_compile_args=['-fopenmp', '-Xlinker', ' -export-dynamic', '-O3', '-std=c++17'],
+            include_dirs=["./parallel_hashmap"],
+        ),
+    ],
+    cmdclass={
+        'build_ext': BuildExtension
+    })

starrygl/sample/stream_manager.py

+import torch
+class WorkStreamEvent:
+    def __init__(self):
+        self.train_stream = torch.cuda.Stream()
+        self.write_memory_stream = torch.cuda.Stream()
+        self.fetch_stream = torch.cuda.Stream()
+        self.write_mail_stream = torch.cuda.Stream()
\ No newline at end of file

train_tgnn.py

+import argparse
+import os
+import sys
+from os.path import abspath, join, dirname
+from starrygl.distributed.context import DistributedContext
+from starrygl.distributed.utils import DistIndex
+from starrygl.module.modules import GeneralModel
+
+from starrygl.module.utils import parse_config
+from starrygl.sample.graph_core import DataSet, GraphData, TemporalNeighborSampleGraph
+from starrygl.sample.memory.shared_mailbox import SharedMailBox
+from starrygl.sample.sample_core.base import NegativeSampling
+from starrygl.sample.sample_core.neighbor_sampler import NeighborSampler
+from starrygl.sample.part_utils.partition_tgnn import partition_load
+import torch
+import time
+import torch
+import torch.nn.functional as F
+import torch.distributed as dist
+import torch.multiprocessing as mp
+from torch.nn.parallel import DistributedDataParallel as DDP
+from torch.distributed import init_process_group, destroy_process_group
+
+import os
+
+from starrygl.sample.data_loader import DistributedDataLoader
+from starrygl.sample.batch_data import SAMPLE_TYPE
+"""
+test command 
+python test.py --world_size 2 --rank 0 
+--world_size', default=4, type=int, metavar='W',
+                    help='number of workers')
+parser.add_argument('--rank', default=0, type=int, metavar='W',
+                    help='rank of the worker')
+parser.add_argument('--log_interval', type=int, default=10, metavar='N',
+                    help='interval between training status logs')
+parser.add_argument('--gamma', type=float, default=0.99, metavar='G',
+                    help='how much to value future rewards')
+parser.add_argument('--seed', type=int, default=1, metavar='S',
+                    help='random seed  for reproducibility')
+parser.add_argument('--num_sampler', type=int, default=10, metavar='S',
+                    help='number of samplers')
+parser.add_argument('--queue_size', type=int, default=10, metavar='S',
+                    help='sampler queue size')
+"""
+parser = argparse.ArgumentParser(
+    description="RPC Reinforcement Learning Example",
+    formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+)
+parser.add_argument('--rank', default=0, type=str, metavar='W',
+                    help='name of dataset')
+parser.add_argument('--world_size', default=1, type=int, metavar='W',
+                    help='number of negative samples')
+args = parser.parse_args()
+from sklearn.metrics import average_precision_score, roc_auc_score
+import torch
+import time
+import random
+import dgl
+import numpy as np
+from sklearn.metrics import average_precision_score, roc_auc_score
+from torch.nn.parallel import DistributedDataParallel as DDP
+os.environ['CUDA_VISIBLE_DEVICES'] = str(args.rank)
+os.environ["RANK"] = str(args.rank)
+os.environ["WORLD_SIZE"] = str(args.world_size)
+os.environ["LOCAL_RANK"] = str(0)
+os.environ["MASTER_ADDR"] = '127.0.0.1'
+os.environ["MASTER_PORT"] = '9337'
+def seed_everything(seed=42):
+    random.seed(seed)
+    np.random.seed(seed)
+    torch.manual_seed(seed)
+    torch.cuda.manual_seed(seed)
+    torch.backends.cudnn.deterministic = True
+    torch.backends.cudnn.benchmark = False
+
+seed_everything(1234)
+def main():   
+    use_cuda = True
+    sample_param, memory_param, gnn_param, train_param = parse_config('./config/TGN.yml')
+    torch.set_num_threads(12)
+    ctx = DistributedContext.init(backend="nccl", use_gpu=True)
+    device_id = torch.cuda.current_device()
+    print('use cuda on',device_id)
+    pdata = partition_load("./dataset/here/WIKI", algo="metis_for_tgnn")    
+    graph = GraphData(pdata = pdata)
+    sample_graph = TemporalNeighborSampleGraph(sample_graph = pdata.sample_graph,mode = 'full')
+    mailbox = SharedMailBox(pdata.ids.shape[0], memory_param, dim_edge_feat = pdata.edge_attr.shape[1] if pdata.edge_attr is not None else 0)
+    sampler = NeighborSampler(num_nodes=graph.num_nodes, num_layers=1, fanout=[10],graph_data=sample_graph, workers=10,policy = 'recent',graph_name = "wiki_train")
+    train_data = torch.masked_select(graph.edge_index,pdata.train_mask.to(graph.edge_index.device)).reshape(2,-1)
+    train_ts = torch.masked_select(graph.edge_ts,pdata.train_mask.to(graph.edge_index.device))
+    val_data = torch.masked_select(graph.edge_index,pdata.val_mask.to(graph.edge_index.device)).reshape(2,-1)
+    val_ts = torch.masked_select(graph.edge_ts,pdata.val_mask.to(graph.edge_index.device))
+    test_data = torch.masked_select(graph.edge_index,pdata.test_mask.to(graph.edge_index.device)).reshape(2,-1)
+    test_ts = torch.masked_select(graph.edge_ts,pdata.test_mask.to(graph.edge_index.device)) 
+    print(train_data.shape[1],val_data.shape[1],test_data.shape[1])
+    train_data = DataSet(edges = train_data,ts =train_ts,eids = torch.nonzero(pdata.train_mask).view(-1))
+    test_data = DataSet(edges = test_data,ts =test_ts,eids = torch.nonzero(pdata.test_mask).view(-1))
+    val_data = DataSet(edges = val_data,ts = val_ts,eids = torch.nonzero(pdata.val_mask).view(-1))
+    #train_neg_sampler = PreNegativeSampling('triplet',torch.masked_select(pdata.edge_index['pos_edge'],graph.data.train_mask).reshape(2,-1))
+    neg_sampler = NegativeSampling('triplet')
+    trainloader = DistributedDataLoader(graph,train_data,sampler = sampler,
+                                        sampler_fn = SAMPLE_TYPE.SAMPLE_FROM_TEMPORAL_EDGES,
+                                        neg_sampler=neg_sampler,
+                                        batch_size = 2000,
+                                        shuffle=False,
+                                        drop_last=True,
+                                        chunk_size = None,
+                                        train=True,
+                                        queue_size = 100,
+                                        mailbox = mailbox)
+    testloader = DistributedDataLoader(graph,test_data,sampler = sampler,
+                                        sampler_fn = SAMPLE_TYPE.SAMPLE_FROM_TEMPORAL_EDGES,
+                                        neg_sampler=neg_sampler,
+                                        batch_size = 2000,
+                                        shuffle=False,
+                                        drop_last=False,
+                                        chunk_size = None,
+                                        train=False,
+                                        queue_size = 100,
+                                        mailbox = mailbox)
+    valloader = DistributedDataLoader(graph,val_data,sampler = sampler,
+                                        sampler_fn = SAMPLE_TYPE.SAMPLE_FROM_TEMPORAL_EDGES,
+                                        neg_sampler=neg_sampler,
+                                        batch_size = 2000,
+                                        shuffle=False,
+                                        drop_last=False,
+                                        chunk_size = None,
+                                        train=False,
+                                        queue_size = 100,
+                                        mailbox = mailbox)
+    gnn_dim_node = 0 if graph.x is None else pdata.x.shape[1]
+    gnn_dim_edge = 0 if graph.edge_attr is None else pdata.edge_attr.shape[1]
+    avg_time  = 0
+    if use_cuda:
+        model = GeneralModel(gnn_dim_node, gnn_dim_edge, sample_param, memory_param, gnn_param, train_param).cuda()
+        device = torch.device('cuda')
+    else:
+        model = GeneralModel(gnn_dim_node, gnn_dim_edge, sample_param, memory_param, gnn_param, train_param)
+        device = torch.device('cpu')
+    model = DDP(model,find_unused_parameters=True)
+    train_stream = torch.cuda.Stream()
+    send_stream = torch.cuda.Stream()
+    scatter_stream = torch.cuda.Stream()
+
+    val_losses = list()
+    def eval(mode='val'):
+        neg_samples = 1
+        model.eval()
+        aps = list()
+        aucs_mrrs = list()
+        if mode == 'val':
+            loader = valloader
+        elif mode == 'test':
+            loader = testloader
+        elif mode == 'train':
+            loader = trainloader
+        with torch.no_grad():
+            total_loss = 0
+            signal = torch.tensor([0],dtype = int,device = device)
+            for roots,mfgs,metadata in loader:
+                signal[0] = 0
+                dist.all_reduce(signal,async_op=False)
+                pred_pos, pred_neg = model(mfgs,metadata)
+                total_loss += creterion(pred_pos, torch.ones_like(pred_pos))
+                total_loss += creterion(pred_neg, torch.zeros_like(pred_neg))
+                y_pred = torch.cat([pred_pos, pred_neg], dim=0).sigmoid().cpu()
+                y_true = torch.cat([torch.ones(pred_pos.size(0)), torch.zeros(pred_neg.size(0))], dim=0)
+                aps.append(average_precision_score(y_true, y_pred.detach().numpy()))
+                aucs_mrrs.append(roc_auc_score(y_true, y_pred))
+                if mailbox is not None:
+                    src = metadata['src_pos_index']
+                    dst = metadata['dst_pos_index']
+                    
+                    ts = roots.ts
+                    if(graph.edge_attr.device == torch.device('cpu')):
+                        edge_feats = graph.edge_attr[roots.eids.to('cpu')].to('cuda') if graph.edge_attr is not None else None
+                    else:
+                        edge_feats = graph.edge_attr[roots.eids] if graph.edge_attr is not None else None
+                    dist_index_mapper = mfgs[0][0].srcdata['ID']
+                    root_index = torch.cat((src,dst))
+                    last_updated_nid = model.module.memory_updater.last_updated_nid[root_index]
+                    last_updated_memory = model.module.memory_updater.last_updated_memory[root_index]
+                    last_updated_ts=model.module.memory_updater.last_updated_ts[root_index]
+                    index, memory, memory_ts = mailbox.get_update_memory(last_updated_nid,
+                                                                       last_updated_memory,
+                                                                       last_updated_ts)
+                    
+                    index, mail, mail_ts = mailbox.get_update_mail(dist_index_mapper,
+                                                 src,dst,ts,edge_feats,
+                                                 model.module.memory_updater.last_updated_memory,
+                                                 )
+                    #mailbox.set_mailbox_all_to_all(index,memory,memory_ts,mail,mail_ts,reduce_Op = 'max')
+            if mode == 'val':
+                val_losses.append(float(total_loss))
+            while(signal[0].item() != dist.get_world_size()):
+                signal[0] = 1
+                dist.all_reduce(signal,async_op=False)
+                if(signal[0].item() == dist.get_world_size()):
+                    break
+                if mailbox is not None:
+                    mailbox.set_mailbox_all_to_all(torch.tensor([],device = device).reshape(-1),
+                                                   torch.tensor([],device = device).reshape(-1,mailbox.memory_size),
+                                                   torch.tensor([],device = device).reshape(-1),
+                                                   torch.tensor([],device = device).reshape(-1,mailbox.mailbox.accessor.data.size(2)),
+                                                   torch.tensor([],device = device).reshape(-1),
+                                                   reduce_Op = 'max')
+
+        ap = float(torch.tensor(aps).mean())
+        if neg_samples > 1:
+            auc_mrr = float(torch.cat(aucs_mrrs).mean())
+        else:
+            auc_mrr = float(torch.tensor(aucs_mrrs).mean())
+        return ap, auc_mrr
+    
+    creterion = torch.nn.BCEWithLogitsLoss()
+    optimizer = torch.optim.Adam(model.parameters(), lr=train_param['lr'])
+    for e in range(train_param['epoch']):
+        epoch_start_time = time.time()
+        train_aps = list()
+        print('Epoch {:d}:'.format(e))
+        time_prep = 0
+        total_loss = 0
+        model.train()
+        if mailbox is not None:
+            mailbox.reset()
+            model.module.memory_updater.last_updated_nid = None
+            model.module.memory_updater.last_updated_memory = None
+            model.module.memory_updater.last_updated_ts = None
+        for roots,mfgs,metadata in trainloader:
+            t_prep_s = time.time()
+            optimizer.zero_grad()
+            with torch.cuda.stream(train_stream):
+                pred_pos, pred_neg = model(mfgs,metadata)
+                loss = creterion(pred_pos, torch.ones_like(pred_pos))
+                loss += creterion(pred_neg, torch.zeros_like(pred_neg))
+                total_loss += float(loss)
+                loss.backward()
+                optimizer.step()
+                t_prep_s = time.time()
+                y_pred = torch.cat([pred_pos, pred_neg], dim=0).sigmoid().cpu()
+                y_true = torch.cat([torch.ones(pred_pos.size(0)), torch.zeros(pred_neg.size(0))], dim=0)
+                train_aps.append(average_precision_score(y_true, y_pred.detach().numpy()))
+                if mailbox is not None:
+                    src = metadata['src_pos_index']
+                    dst = metadata['dst_pos_index']
+                    
+                    ts = roots.ts
+                    if(graph.edge_attr.device == torch.device('cpu')):
+                        edge_feats = graph.edge_attr[roots.eids.to('cpu')].to('cuda') if graph.edge_attr is not None else None
+                    else:
+                        edge_feats = graph.edge_attr[roots.eids] if graph.edge_attr is not None else None
+                    dist_index_mapper = mfgs[0][0].srcdata['ID']
+                    root_index = torch.cat((src,dst))
+                    last_updated_nid = model.module.memory_updater.last_updated_nid[root_index]
+                    last_updated_memory = model.module.memory_updater.last_updated_memory[root_index]
+                    last_updated_ts=model.module.memory_updater.last_updated_ts[root_index]
+                    index, memory, memory_ts = mailbox.get_update_memory(last_updated_nid,
+                                                                       last_updated_memory,
+                                                                       last_updated_ts)
+                    index, mail, mail_ts = mailbox.get_update_mail(dist_index_mapper,
+                                                 src,dst,ts,edge_feats,
+                                                 model.module.memory_updater.last_updated_memory,
+                                                 )
+                    #mailbox.set_mailbox_all_to_all(index,memory,memory_ts,mail,mail_ts,reduce_Op = 'max')
+        torch.cuda.synchronize()
+        time_prep = time.time() - epoch_start_time
+        avg_time += time.time() - epoch_start_time
+        train_ap = float(torch.tensor(train_aps).mean())    
+        ap = 0
+        auc = 0
+        ap, auc = eval('val')
+        print('\ttrain loss:{:.4f}  train ap:{:4f}  val ap:{:4f}  val auc:{:4f}'.format(total_loss,train_ap, ap, auc))
+        print('\ttotal time:{:.2f}s  prep time:{:.2f}s'.format(time.time()-epoch_start_time, time_prep))    
+          
+    model.eval()
+    if mailbox is not None:
+        mailbox.reset()
+        model.module.memory_updater.last_updated_nid = None
+        eval('train')
+        eval('val')
+    ap, auc = eval('test')
+    eval_neg_samples = 1
+    if eval_neg_samples > 1:
+        print('\ttest AP:{:4f}  test MRR:{:4f}'.format(ap, auc))
+    else:
+        print('\ttest AP:{:4f}  test AUC:{:4f}'.format(ap, auc))    
+    print('test_dataset',test_data.edges.shape[1],'avg_time',avg_time/train_param['epoch'])
+    ctx.shutdown()
+if __name__ == "__main__":
+    main()
+