文件历史

提交图

73 次代码提交

作者 SHA1 备注 提交日期
Jinjing Zhou cd204a4ac9 [Distributed] Use barrier instead of sleep in DistDataloader (#2086)
* use barrier instead of sleep

* lint
2020-08-21 18:01:09 +08:00
Chao Ma a4a7e02f5f [BugFix] Fix a small bug on kvstore (#2058)
* small fix

* update
2020-08-19 14:57:45 +08:00
Da Zheng 9da1629c9d [Distributed] Some small fixes on doc (#2049)
* quick fix.

* update sparse optimizer.

* fix.

* fix
2020-08-18 21:13:19 -07:00
Da Zheng 75ffc31f76 [Doc] Update the docstring of distributed APIs. (#2025)
* add doc.

* update DistGraph.

* add DistTensor.

* update DistEmbedding.

* add partition.py

* add sampling.

* fix.

* add graph partition book and create a base class.

* fix test.

* add rst.

* update doc rst.

* update.

* fix.

* fix docs

* update distributed tensor and embeddings.

* add checks.

* update DistGraph.

* update initialization.

* fix graph partition book.

* update graph partition book.

* update partition.

* update partition.

* fix.

* add example code.

* update DistGraph

* Update python/dgl/distributed/dist_context.py

Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>

* Update python/dgl/distributed/dist_context.py

Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>

* Update python/dgl/distributed/dist_dataloader.py

Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>

* Update python/dgl/distributed/dist_dataloader.py

Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>

* Update python/dgl/distributed/dist_dataloader.py

Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>

* update initialize.

* update dataloader.

* update distgraph.

* update DistGraph.

* update DistTensor.

* update.

* more updates.

* fix lint.

* add num_nodes and num_edges

Co-authored-by: Chao Ma <mctt90@gmail.com>
Co-authored-by: Quan (Andy) Gan <coin2028@hotmail.com>
Co-authored-by: xiang song(charlie.song) <classicxsong@gmail.com>
2020-08-18 11:53:45 -07:00
Quan (Andy) Gan edfbee2ca8 [Feature] Make sample_neighbors copy features on demand (#2042)
* fix

* fix

* lint

* fix

* test

* fix

* fix
2020-08-18 22:25:19 +08:00
Da Zheng efa95d8673 [Distributed] Fix bugs (#2009)
* fix.

* add tests.

* fix.

* fix.
2020-08-13 23:51:11 -07:00
Da Zheng 1ad46fd015 [Distributed] support distributed sampling in NodeDataLoader (#2012)
* fix.

* doesn't support EdgeDataLoader

* add test.

* fix.

* fix test.

* catch errors in the test.

* fix.

* fix error message

* add print.

* join the server processes.

* update test.

* fix test.
2020-08-13 22:49:21 -07:00
Chao Ma 15e3ff878c [Distributed] Fix memory leak in kvstore (#2016)
* Fix memory leak

* update

* update

* update

* update

* update

* update

* update
2020-08-13 17:17:24 +08:00
Da Zheng d1cf5c383e [Distributed] adjust various APIs. (#1993)
* rename get_data_size.

* remove g from DistTensor.

* remove g from DistEmbedding.

* clean up API of graph partition book.

* fix DistGraph

* fix lint.

* collect all part policies.

* fix.

* fix.

* support distributed sampler.

* remove partition.py
2020-08-11 20:22:00 -07:00
Da Zheng 4efa332044 [Distributed] Move server start code to initialize. (#2002)
* move server start code to initialize.

* fix.

* fix lint.

* fix examples.

* add more checks.
2020-08-11 16:58:18 -07:00
Chao Ma d340ea3aa8 [Distributed] Remove server_count from ip_config.txt (#1985)
* remove server_count from ip_config.txt

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* lint

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* Update dist_context.py

* fix lint.

* make it work for multiple spaces.

* update ip_config.txt.

* fix examples.

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* udpate

* update

* update

* update

* update

* update

Co-authored-by: Da Zheng <zhengda1936@gmail.com>
Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
2020-08-11 17:41:48 +08:00
chwan-rice 71f4230a8b [Fix] OpenMP compatibility issues (#1992)
* fix OpenMP compatibility issues

* typo

Co-authored-by: Jinjing Zhou <VoVAllen@users.noreply.github.com>
2020-08-11 14:26:34 +08:00
Da Zheng ee30b2aadd [Distributed] Fix standalone (#1974)
* fix tests.

* fix.

* remove a test.

* make code work in the standalone mode.

* fix example.

* more fix.

* make DistDataloader work with num_workers=0

* fix DistDataloader tests.

* fix.

* fix lint.

* fix cleanup.

* fix test

* remove unnecessary code.

* remove tests.

* fix.

* fix.

* fix.

* fix example

* fix.

* fix.

* fix launch script.

Co-authored-by: Jinjing Zhou <VoVAllen@users.noreply.github.com>
Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
2020-08-10 15:49:28 -07:00
Da Zheng b9ef70e5ac [Distributed] Set the number of threads correctly to speed up (#1976)
* temp fix omp.

* set server threads.

* add CAPI to set up OMP threads.

* fix.

* fix.

* update namesapce.

* set cpi properly.

* allow to config num worker threads.

* set #threads.

* fix.
2020-08-08 19:30:53 -07:00
Da Zheng 5454471fc8 [Distributed] Add roles (#1971)
* distinguish roles.

* add comments.

* fix lint.

* move roles to server_state.

* fix text.

* fix tests.

* fix tests.

* Revert "fix tests."

This reverts commit 5baa136b872a4550d4e612bfb1dfe363d7814adf.
2020-08-08 01:41:51 -07:00
Da Zheng 2823c61f37 [Distributed] Fix synchronized APIs in kvstore (#1948)
* handle synchronized API.

* fix.

* fix.

* fix.

Co-authored-by: Chao Ma <mctt90@gmail.com>
2020-08-07 19:01:52 -07:00
Da Zheng bab32d5ba1 [Distributed] Fix dataloader (#1970)
* fix dataloader.

* initialize iterator of DistDataloader correctly.

* update test.
2020-08-07 18:09:21 -07:00
Minjie Wang 22167f7203 [Refactor] Enable new kernel in all message passing APIs (#1953)
* WIP: frame refactor

* new frame

* simple update_all builtin

* move all subgraph routines into the same file

* sddmm & spmm schedule; node & edge udf

* degree bucketing

* some tricky 0deg corner cases

* bug in frame append

* merge test_hetero_basics and test_basics

* some code rearange

* fix test_heterograph

* add mean spmm

* enable all builtin combinations

* pass gpu test

* pass pytorch tests

* wip

* fix some pt debugging codes

* fix bug in mxnet backward

* pass all mxnet utests

* passed tf tests

* docstring

* lint

* lint

* fix broadcasting bugs

* add warning and clamp for mean reducer

* add test for zero-degree mean

* address comments

* lint

* small fix
2020-08-07 15:40:25 +08:00
Jinjing Zhou 4f499c7ffb [Distributed] DistDataloader (#1901)
* 111

* 111

* fix

* 111

* fix

* 11

* fix

* lint

* Update __init__.py

* lint

* fix

* lint

* fix

* fix

* fix

* fix

* fix

* try fix

* try fix

* fix

* Revert "fix"

This reverts commit a0b954fd4e99b7df92b53db8334dcb583d6e1551.

* fixes.

* fix.

* fix test.

* fix exit.

* fix.

* fix

* fix

* lint

* lint

* lint

* fix

* Update .gitignore

* 111

* fix

* 111

* 111

* fff

* 1111

* 111

* 1325315

* ffff

* f???

* fff

* 1111

* 111

* fix

* 111

* asda

* 1111

* 11

* 123

* 啊啊啊啊啊啊啊啊啊啊啊啊啊啊啊啊

* spawn

* 1231231

* up

* 111

* fix

* fix

* Revert "fix"

This reverts commit 7373f95312fdcaa36d2fc330bf242339e89c045d.

* fix

* fix

* 1111

* fix

* fix tests

* start kvclient as early as possible.

* lint

* fix test

* lint

* 1111

* fix

* fix

* 111

* fix

* fix

* 1

* fix

* fix

* lint

* fix

* lint

* lint

* remove quit

* fix

* lint

* fix

* fix several

* lint

* fix minor

* fix

* lint

Co-authored-by: Da Zheng <zhengda1936@gmail.com>
Co-authored-by: xiang song(charlie.song) <classicxsong@gmail.com>
2020-08-05 17:43:17 +08:00
Da Zheng a4c931a9bf [Distributed] Support multiple servers (#1886)
* client init graph on the backup servers.

* fix.

* test multi-server.

* fix anonymous dist tensors.

* check #parts.

* fix init_data

* add multi-server multi-client tests.

* update tests in kvstore.

* fix.

* verify the loaded partition.

* fix a bug.

* fix lint.

* fix.

* fix example.

* fix rpc.

* fix pull/push handler for backup kvstore

* fix example readme.

* change ip.

* update docstring.

Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
2020-08-02 19:50:34 -07:00
xiang song(charlie.song) 56bbf9cb6d [Distributed][Example] Distributed training for unsupervised graphsage (#1853)
* Standalone can run

* fix

* Add save

* Fix

* Fix

* Fix

* Fix

* debug

* test

* test

* Fix

* Fix

* log

* Fix

* fix

* Profile

* auto sync grad

* update

* add test for unsupervised dist training

* upd

* Fix lr

* Fix update

* sync

* fix

* Revert "fix"

This reverts commit d5caa7398b36125f6d6e2c742a95c6ff4298c9e9.

* Fix

* unsupervised

* Fix

* remove debug

* Add test case for dist_graph find_edges()

* Fix

* skip tensorflow test for find_edges

* Update readme

* remove some test

* upd

* Update partition_graph.py

Co-authored-by: Ubuntu <ubuntu@ip-172-31-68-185.ec2.internal>
Co-authored-by: Ubuntu <ubuntu@ip-172-31-24-210.ec2.internal>
Co-authored-by: Da Zheng <zhengda1936@gmail.com>
2020-08-01 12:24:55 -07:00
Da Zheng 4be4b13424 [Distributed] add copy_partitions.py (#1866)
* fix bugs.

* eval on both vaidation and testing.

* add script.

* update.

* update launch.

* make train_dist.py independent.

* update readme.

* update readme.

* update readme.

* update readme.

* generate undirected graph.

* rename conf_file to part_config

* use rsync

* make train_dist independent.

Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-115.us-west-2.compute.internal>
Co-authored-by: xiang song(charlie.song) <classicxsong@gmail.com>
2020-07-31 11:11:56 -07:00
Chao Ma e5e1a99d46 [KVStore] Fix bug in barrier. (#1904)
* update

* update

* fix lint

* update

* update
2020-07-31 22:01:03 +08:00
Chao Ma 5b515cf667 [KVStore] Support group barrier (#1880)
* udpate

* update

* update

* update

* update

* update

* update

* update

* fix lint

* update

* update

* update

* update

* udpate

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

Co-authored-by: Da Zheng <zhengda1936@gmail.com>
2020-07-30 19:21:20 +08:00
Da Zheng 3e2f94ed20 [Distributed] Refactor distributed training to new DGLGraph. (#1874)
* fix tests in partition.

* fix DistGraph.

* fix without shared memory.

* fix sampling.

* enable distributed test.

* fix tests.

* fix a bug in shared-mem heterograph.

* print better error messages.

* fix.

* don't specify formats.

* fix.

* fix

* small fix.
2020-07-28 22:07:24 -07:00
Jinjing Zhou 71cf78652d Revert "[Distributed] DistDataloader (#1870)" (#1876)
This reverts commit 6557291fdc.
2020-07-28 17:57:57 +08:00
Jinjing Zhou 6557291fdc [Distributed] DistDataloader (#1870)
* 111

* 111

* fix

* 111

* fix

* 11

* fix

* lint

* Update __init__.py

* lint

* fix

* lint

* fix

* fix

* fix

* fix

* fix

* try fix

* try fix

* fix

* Revert "fix"

This reverts commit a0b954fd4e99b7df92b53db8334dcb583d6e1551.

* fixes.

* fix.

* fix test.

* fix exit.

* fix.

* fix

* fix

* lint

* lint

* lint

* fix

* Update .gitignore

Co-authored-by: Da Zheng <zhengda1936@gmail.com>
2020-07-28 16:36:37 +08:00
Da Zheng 015acfd2d8 [Distributed] Small fixes (#1859)
* fix node/edge_split.

* fix partition.

* support heterograph interface.

* fix test.

* fix

* fix docstring.
2020-07-27 01:39:48 -07:00
Da Zheng ff8f7082a3 [Distributed] turn off recording on embeddings in the inference. (#1861)
* turn on/off recording in sparse embedding.

* add test.
2020-07-27 00:42:40 -07:00
Da Zheng 6963d796c5 [Distributed] update Embedding API (#1847)
* update.

* update Embedding.

* add comments.

* fix lint

Co-authored-by: Chao Ma <mctt90@gmail.com>
2020-07-26 11:59:41 -07:00
Da Zheng 562871e76b [Distributed] add distributed evaluation. (#1810)
* add eval.

* extend DistTensor.

* fix.

* add barrier.

* add more print.

* add more checks in kvstore.

* fix lint.

* get all neighbors for eval.

* reorganize.

* fix.

* fix.

* fix.

* fix test.

* add reuse_if_exist.

* add test for reuse_if_exist.

* fix lint.

* fix bugs.

* fix.

* print errors of tcp socket.

* support delete tensors.

* fix lint.

* fix

* fix example

Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
2020-07-22 00:40:14 -07:00
Chao Ma 5c92f6c218 [RPC] Rpc exit with explicit invocation (#1825)
* exit client

* update

* update

* update

* update

* update

* update

* update

* update test

* update

* update

* update

* update

* update

* update

* update

* update

* update
2020-07-20 14:56:26 +08:00
Chao Ma 65e7805e22 [KVStore] Support delete_data() api on client (#1824)
* delete tensor

* update

* update

* update

* update

* update

* update

* udpate

* update

* update

* update
2020-07-20 12:24:23 +08:00
Da Zheng ea420c0abd [Distributed] Reduce memory consumption in graph partitioning (#1823)
* save mem.

* save mem.

* reduce mem

* fix test

* fix lint

* fix test

* fix.

* fix.

* fix.

* fix.

* fix lint.

* fix backend operator.

* fix tensorflow operators.

* fix.

* revert change in mxnet operator.

Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
2020-07-18 06:53:48 -07:00
Da Zheng 20734637f0 [Distributed] Fix partition (#1821)
* fix partition and print timing.

* fix lint.

Co-authored-by: Ubuntu <ubuntu@ip-172-31-19-1.us-west-2.compute.internal>
2020-07-17 16:03:37 -07:00
Da Zheng cda0abf75a [Distributed] add the standalone mode in DistGraph (#1800)
* add standalone mode

* add comments.

* add tests for sampling.

* fix.

* make the code to run the standalone mode

* fix

* fix

* fix readme.

* fix.

* fix test

Co-authored-by: Chao Ma <mctt90@gmail.com>
2020-07-14 23:46:25 -07:00
Da Zheng a1472bcff6 [Distributed] Run distributed graph server inside DGL (#1801)
* run dist server in dgl.

* fix bugs.

* fix example.

* check environment variables and fix lint.

* fix lint
2020-07-14 07:41:44 -07:00
Da Zheng 6a3685bea0 [BUGFIX] fix a bug in graph partitioning. (#1769)
* fix

* use utils.toindex in the right place.

* fix.

* update tensor for mxnet backend.

* fix

* fix
2020-07-09 01:16:42 -07:00
Da Zheng 167216afc8 [Distributed] add in_subgraph on DistGraph. (#1755)
* add in_subgraph on DistGraph.

* check in more.

* fix test.

* add comments.

* fix test.

* update test.

* update.

* rename.

* update comment

* fix test
2020-07-08 14:39:21 -07:00
Da Zheng afb7f5b839 [Distributed] fix init_data in kvstore. (#1750)
* fix

* fix.

* update.

* fix

* add assert

Co-authored-by: Chao Ma <mctt90@gmail.com>
2020-07-06 23:02:00 +08:00
Da Zheng cadcc1c2a6 [Feature] add sparse embedding. (#1497)
* add sparse embedding.

* fix

* add test.

* man fixes.

* many fixes

* fix sparse emb.

* fix.

* fix lint.

* fix lint.

* fix kvstore.

* expose DistTensor.

* test sparse embeddings.

* add attach_grad to the backends.

* remove part_id

* fix.

* move backward computation.

* move more computation to backend.

* fix a bug when applying learning rate.

* fix a few things.

* fix a few things.

* add docstring

* fix.

* apply no_grad.

* fix tests.

* fix for other frameworks.

* add examples in docstring.
2020-07-03 10:42:09 -07:00
Chao Ma 781b9cec4c [Fix] Add RangePartitionBook to __init__ (#1738)
* update

* update
2020-07-02 16:05:32 +08:00
Chao Ma 9eb0efcf2a [KVStore] Test multi-server and multi-client (#1732)
* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update

* update
2020-07-01 20:26:40 -07:00
Da Zheng 2c88f7c0b0 [Distributed] Split data to ensure data locality for multiple clients (#1710)
* fix.

* fix tests.

* fix

* add tests.

* fix.

* have default rank.

* add comment.

* fix test.

* remove check

* simplify code.

* add test.

* split data evenly.

* simplify the distributed training code.

* add comments.

* add comments.
2020-07-01 13:03:40 -07:00
mszarma 69e6afdf82 Fix crashing on SIOCGIFADDR (#1733)
During scanning the list of interfaces,
some of them can be configured invalidly:
lack of ip, link down etc.
This patch handles the error
and inform about problem with concrete interface.
2020-07-01 22:47:24 +08:00
Jinjing Zhou 18a26fcfb1 [Data] Support HeteroGraph save/load (#1526)
* 111

* add history version test

* fix

* 111

* save

* ``

* fix1

* 111

* add save heterograph

* lint

* lint

* add tests

* minor fix

* fix

* docs

* add format tets

* use unique_ptr

* fix

* fix interface

* 111

* 111

* fix

* lint

* fix

* add support to s3

* fix

* fix

* fix leak

* fix

* fix docs

* fix

* linlt

* fix

* fix

* fix

* address comment

* address comment

Co-authored-by: Minjie Wang <wmjlyjemaine@gmail.com>
Co-authored-by: xiang song(charlie.song) <classicxsong@gmail.com>
2020-06-30 13:34:50 +08:00
Chao Ma dd8d5289ef [rpc] Add num_clients and sort client ID by machine-id (#1713)
* add num_clients to kvstore

* update

* update

* update

* update

* update

* update

* update

* fix lint

* update

* update

* update

* add test

* update

* update

* update

* update

* fix test

* update

* update

* update

* update

* update

* update

* update

Co-authored-by: Da Zheng <zhengda1936@gmail.com>
2020-06-28 18:40:38 -07:00
Da Zheng 02d3197407 [Distributed] Pytorch example of distributed GraphSage. (#1495)
* add train_dist.

* Fix sampling example.

* use distributed sampler.

* fix a bug in DistTensor.

* fix distributed training example.

* add graph partition.

* add command

* disable pytorch parallel.

* shutdown correctly.

* load diff graphs.

* add ip_config.txt.

* record timing for each step.

* use ogb

* add profiler.

* fix a bug.

* add train_dist.

* Fix sampling example.

* use distributed sampler.

* fix a bug in DistTensor.

* fix distributed training example.

* add graph partition.

* add command

* disable pytorch parallel.

* shutdown correctly.

* load diff graphs.

* add ip_config.txt.

* record timing for each step.

* use ogb

* add profiler.

* add Ips of the cluster.

* fix exit.

* support multiple clients.

* balance node types and edges.

* move code.

* remove run.sh

* Revert "support multiple clients."

* fix.

* update train_sampling.

* fix.

* fix

* remove run.sh

* update readme.

* update readme.

* use pytorch distributed.

* ensure all trainers run the same number of steps.

* Update README.md

Co-authored-by: Ubuntu <ubuntu@ip-172-31-16-250.us-west-2.compute.internal>
2020-06-28 13:09:40 +08:00
Da Zheng 04d4680d5b [Distributed] add constraints for metis partitioning. (#1708)
* add constraint.

* fix bugs.

* temp fix.

* fix lint

* add balance_ntypes and balance_edges

* add comments.

* fix.

* fix
2020-06-27 21:27:18 -07:00
Da Zheng 45e1333e13 [Distributed] Optimize distributed sampling (#1700)
* optimize sampling neighbors from the local partition.

* fix minior bugs.

* fix lint

* overlap local sampling and remote sampling.

* fix.

* fix lint.

* fix a potential bug.

Co-authored-by: Chao Ma <mctt90@gmail.com>
2020-06-26 17:05:29 -07:00