summaryrefslogtreecommitdiff
path: root/torch/lib
AgeCommit message (Expand)AuthorFilesLines
2019-02-05Rename IntList to IntArrayRef. (#16751)Edward Yang1-5/+5
2019-01-30Include ATen/core/functional.h directly instead of torch/csrc/utils/functiona...Mikhail Zolotukhin2-2/+2
2019-01-29try to get rid of tmp_install (#16414)Zachary DeVito4-12/+4
2019-01-29Make Store::setTimeout take milliseconds (#16278)Pieter Noordhuis2-4/+4
2019-01-24Move THCCachingAllocator to c10_cuda. (#16119)Edward Yang2-8/+8
2019-01-24Remove unnecessary includes and headers from THCCachingAllocator, move to at:...Edward Yang2-8/+8
2019-01-23CUDA event should only be recorded after NCCL group (#8219)Teng Li1-12/+40
2019-01-18TCP init method race condition fix (#15684)Teng Li3-7/+74
2019-01-16Generate PDB files for better debugging on Windows (#16008)peter1-0/+4
2019-01-14Fix c10d checking errno unconditionally (#15986)SsnL3-27/+43
2019-01-14libshm retry on EINTR (#15964)SsnL4-20/+40
2019-01-08Revert "remove use of tmp_install" (#15847)andersj4-4/+4
2019-01-04remove use of tmp_installandersj4-4/+4
2018-12-12Move CUDAGuard, CUDAStream and CUDAGuardImpl to c10/cuda (#14248)Edward Yang7-9/+9
2018-12-11add gloo support for gather on GPU (#14916)Jane Wang1-4/+114
2018-12-11add gloo scatter support on GPU (#14917)Jane Wang1-4/+107
2018-12-10add gloo allgather support on GPU (#14576)Jane Wang1-18/+163
2018-12-08Canonicalize all includes in PyTorch. (#14849)Edward Yang54-102/+92
2018-12-04c10d doesn't work with torch namespace (#14042)Zeming Lin1-1/+1
2018-12-03Fix multi-argument allreduce in ProcessGroupGloo (#14688)Pieter Noordhuis1-1/+13
2018-11-29Enable tests for CPU tensors in test_distributed.py (#14572)Pieter Noordhuis1-9/+8
2018-11-29Fixed new_group won't work for two or more different rank groups (#14529)Teng Li2-14/+63
2018-11-29add gloo support for reduce on GPU (#14443)Jane Wang1-11/+100
2018-11-29Add sourceRank() to ProcessGroup::Work (#14453)Pieter Noordhuis9-33/+34
2018-11-27Barrier synchronizes with prior work before completing (#14386)Pieter Noordhuis2-36/+76
2018-11-27Make ProcessGroup::Work::wait() throw (#14298)Pieter Noordhuis11-326/+145
2018-11-27Add option structs and timeout field (#14297)Pieter Noordhuis8-16/+46
2018-11-27Refer to all work with ProcessGroup prefix (#14296)Pieter Noordhuis2-12/+12
2018-11-27Remove algorithm caching in ProcessGroupGloo (#14295)Pieter Noordhuis3-602/+24
2018-11-27Use new style barrier support in c10d/gloo (#14294)Pieter Noordhuis1-6/+23
2018-11-26Fixed file init_method write/read race (#14388)Teng Li1-6/+24
2018-11-21Use ADL to find toString (#14021)Sebastian Messmer1-2/+2
2018-11-21Robust NCCL barrier improvement to cover all devices combinations (#14271)Teng Li2-6/+10
2018-11-21Remove header generated at configuration time (#14244)Pieter Noordhuis3-14/+12
2018-11-20c10d Automatically retry on EINTR (#14180)Tongzhou Wang1-3/+5
2018-11-20Make NCCL backend support barrier op (#14142)Teng Li2-6/+53
2018-11-19Split out CUDAMultiStreamGuard from CUDAGuard (#13912)Edward Yang2-0/+2
2018-11-18No more -werror for c10d (#14155)Teng Li1-1/+0
2018-11-16Fixed MPI build with higher version of GCC (#14122)Teng Li1-2/+5
2018-11-16add torch-python target (#12742)Anders Papitto2-18/+6
2018-11-14Build with -Werror (#13998)Teng Li3-4/+7
2018-11-14Fixed uninitialized warning (#14001)Teng Li1-1/+1
2018-11-14Canonicalize THD includes with .. in themEdward Yang28-42/+42
2018-11-14Remove defunct build.sh/THConfig.cmake (#13953)Edward Yang1-28/+0
2018-11-14FileStore auto deletes file and FileStore::add bug fix (#13708)Teng Li7-21/+61
2018-11-11New implementations of DeviceGuard, StreamGuard and MultiStreamGuard (with CU...Edward Yang6-74/+58
2018-11-10Fix race condition in TCPStoreDaemon initialization (#13815)Pieter Noordhuis1-7/+6
2018-11-07Kill c10d/private/CUDAUtils.hpp (#13681)Teng Li7-35/+3
2018-11-07codemod tensor.type().is_cuda(), tensor.type().is_sparse() (#13590)Richard Zou2-4/+4
2018-11-06Fixed NCCL backend not being built (#13653)Teng Li1-0/+1
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-25 23:23:54 +0000