summaryrefslogtreecommitdiff
path: root/torch/csrc/cuda
AgeCommit message (Expand)AuthorFilesLines
2018-03-30Use THC cached CUDA device property when get_device_name and get_device_capab...Tongzhou Wang1-26/+0
2018-03-10Minor improvement in AutoGPU usage in CUDA bindings (#5689)Adam Paszke1-1/+1
2018-03-08add gpu guard for broadcast_coalesce (#5655)Ailing1-0/+2
2018-03-05Some additional clean-ups (#5505)Sam Gross1-0/+1
2018-03-01Delete unused files (#5500)Sam Gross1-1/+0
2018-02-28better solution for the linking error related to lazy_init for MSVC (#5375)peterjc1232-32/+0
2018-02-27Delete dead Tensor code paths (#5417)Sam Gross8-289/+59
2018-02-27DataParallel: GPU imbalance warning (#5376)Carl Lemaire1-1/+25
2018-02-26Remove two uses of the old Tensor class (#5413)Sam Gross1-9/+15
2018-02-23Merge Variable and Tensor classes (#5225)Sam Gross2-62/+0
2018-02-21Various dtype improvements. (#5321)gchanan2-0/+32
2018-02-06Check that indices and values are on the same device (#5089)Sam Gross1-0/+1
2018-01-19Fix Python docs for broadcast and braodcast_coalesced (#4727)Sam Gross1-1/+1
2018-01-18Allow Variables in calls to NCCL bindings. (#4725)Sam Gross1-10/+29
2018-01-18Define CHECK in torch/csrc/cuda/nccl.h (#4721)Sam Gross3-14/+14
2018-01-18Move broadcast and broadcast_coalesced to C++Adam Paszke6-0/+162
2018-01-18Base for pure C++ NCCL interfaceAdam Paszke6-351/+437
2018-01-09Methods for checking CUDA memory usage (#4511)Tongzhou Wang1-0/+44
2017-12-20Add build support for Python 2.7 using MSVC (#4226)peterjc1231-3/+3
2017-12-06Implement Variable.from_numpy (#4043)Sam Gross1-1/+2
2017-12-04Add streams and comms as optional arguments (#3968)Sam Gross3-168/+273
2017-11-20Implement Variable.storage() (#3765)Sam Gross2-9/+10
2017-11-16Add cudaEvent support to the profiler (#3734)Zachary DeVito1-0/+41
2017-11-15[ATen] Rename isSparse -> is_sparse.Gregory Chanan1-2/+2
2017-11-15[ATen] Rename isCuda -> is_cuda.Gregory Chanan1-2/+2
2017-11-10added #define __STDC_FORMAT_MACROS to tensor and storage code templates to av...andreh72-0/+4
2017-11-09add warnings if device capability is less than ideal (#3601)Soumith Chintala1-0/+20
2017-11-08Improve Windows Compatibility (for csrc/scripts) (#2941)peterjc1233-6/+7
2017-11-07Exposing emptyCache from allocator (#3518)SsnL1-0/+10
2017-10-13skip ncclCommDestroy if CUDA driver is already unloadedsoumith1-2/+9
2017-10-11Fix macOS build (with CUDA) (#3071)Adam Paszke1-0/+2
2017-10-09implement all_reduce, broadcast, all_gather, reduce_scatterSoumith Chintala3-5/+182
2017-10-09add initial NCCL C bindingsSoumith Chintala3-0/+187
2017-09-30refactor THCP method defs into cuda/Module.cppSoumith Chintala1-0/+30
2017-08-26Add torch.cuda.get_device_name function (#2540)Justin Johnson2-0/+14
2017-08-25Fix typos.Zhou Mo1-1/+1
2017-08-10Revert "Fix typos."Gregory Chanan1-1/+1
2017-08-08Fix typos.Zhou Mo1-1/+1
2017-07-21Fixes for DistributedDataParallel (#2168)Adam Paszke1-3/+24
2017-07-21Warning squash.Edward Z. Yang1-1/+1
2017-07-11initialize aten and pytorch to share the same THCStateZach DeVito1-4/+2
2017-06-13Make csrc -Werror clean. (#1795)Edward Z. Yang1-0/+1
2017-06-12Add DistributedDataParallelAdam Paszke1-1/+4
2017-06-11Move expand_utils-inl.h to generic/ and generate via macros.Gregory Chanan1-183/+4
2017-06-11Clarify a number of comments.Gregory Chanan1-1/+4
2017-06-11Add broadcasting support for copy_, simplify code generation by moving a lot ...Gregory Chanan3-0/+190
2017-05-25make THPPointer have explicit constructors (#1636)Trevor Killeen2-2/+2
2017-04-29Parallelize TensorMethods.cpp builds (#1400)Adam Paszke1-6/+5
2017-04-28Refactor Python string utility functionSam Gross1-13/+0
2017-04-28Revert "Parallelize TensorMethods.cpp builds (#1364)" (#1390)Soumith Chintala1-5/+6
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-30 23:53:00 +0000