OSSForks/tensorflow
1
0
Fork 0
mirror of https://github.com/zebrajr/tensorflow.git synced 2026-01-15 12:15:41 +00:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

Commit Graph

  • 1a5402d93f [XLA:GPU] don't stop traversal when sinking bitcasts Mikhail Goncharov 2025-12-18 10:00:19 -08:00
  • c549ee47f8 [XLA:GPU] Use StreamState as rendezvous value. Oleg Shyshkov 2025-12-18 09:59:09 -08:00
  • 5db58f8f58 [xla:cpu] Do not expand convolution feature group if the convolution is supported by libraries. Penporn Koanantakool 2025-12-18 09:22:30 -08:00
  • 91088251a0 PR #35463: [xla:gpu] Support ncclAlltoall directly for contiguous send/recv buffers Eugene Zhulenev 2025-12-18 08:19:37 -08:00
  • f4c5fe5509 In FileDescriptor tests, improve temporary file path generation. Quentin Khan 2025-12-18 08:14:19 -08:00
  • f644aa87f7 PR #35482: Correctly generate compile_commands.json Eugene Zhulenev 2025-12-18 07:57:35 -08:00
  • fbfba09a1b Remove unnecessary local_defines and add missing includes in gpu_executor_test. Henning Becker 2025-12-18 07:57:11 -08:00
  • 434dd85854 Apply llvm-use-new-mlir-op-builder fixes A. Unique TensorFlower 2025-12-18 07:54:31 -08:00
  • 5a0f4aee01 PR #35510: [ROCm] Initialze collectives to nullptr to force its allocation later Harsha H S 2025-12-18 07:48:53 -08:00
  • a59ffc09dd Reverts 408bf09796 A. Unique TensorFlower 2025-12-18 07:23:17 -08:00
  • 9d0d22dbaa [PJRT] Change the two optimizations in Transpose to operate on Loop nests, rather than on the original dimensions. Peter Hawkins 2025-12-18 07:10:07 -08:00
  • 4baf73c4e8 Refactor: Extract AotCompilationResult to CompiledModule. Henning Becker 2025-12-18 06:48:13 -08:00
  • 50c19ba022 Apply llvm-use-new-mlir-op-builder fixes A. Unique TensorFlower 2025-12-18 06:47:02 -08:00
  • 4e34cc6fb7 [XLA:GPU] Support partitioned across replicas modules A. Unique TensorFlower 2025-12-18 06:28:57 -08:00
  • 5e49ee5ed3 Automated Code Change A. Unique TensorFlower 2025-12-18 06:15:09 -08:00
  • 6457884a9b Automated Code Change A. Unique TensorFlower 2025-12-18 06:07:02 -08:00
  • 17fa72acde Automated Code Change A. Unique TensorFlower 2025-12-18 05:50:44 -08:00
  • b64c84f2c3 Remove forgotten ROCM version checks from NcclCollectives Henning Becker 2025-12-18 05:50:11 -08:00
  • 6286fccd8f Automated Code Change A. Unique TensorFlower 2025-12-18 05:33:23 -08:00
  • d3933721c1 When opening a file, check that the file path is not null. Quentin Khan 2025-12-18 05:31:45 -08:00
  • d5820b3000 Automated Code Change A. Unique TensorFlower 2025-12-18 05:20:19 -08:00
  • f17984d352 Automated Code Change A. Unique TensorFlower 2025-12-18 05:10:43 -08:00
  • 08d6df5eea Update XNNPack version Quentin Khan 2025-12-18 05:06:04 -08:00
  • 69cd9be899 Add a function to check for empty/non existing files. Quentin Khan 2025-12-18 04:47:10 -08:00
  • 6d5546d597 Refactor: Dynamically register custom call targets in custom_call_test.cc Henning Becker 2025-12-18 04:13:39 -08:00
  • 408bf09796 [XLA:GPU]Disable TransposeDimensionGrouper pass and replace it with OTF normalization in emitters Theotime Combes 2025-12-18 03:59:57 -08:00
  • 9024ef1e4c Automated Code Change A. Unique TensorFlower 2025-12-18 03:03:08 -08:00
  • 35808079a2 Automated Code Change A. Unique TensorFlower 2025-12-18 02:58:06 -08:00
  • 90f6a02276 [XLA:GPU] Add method for printing unsatisfied Constraints for ConstraintExpression. Greg Olechwierowicz 2025-12-18 02:48:01 -08:00
  • fe216f0f45 Automated Code Change A. Unique TensorFlower 2025-12-18 02:44:57 -08:00
  • 69f8ca2e28 PR #35479: Add clangd files and directories to .gitignore Eugene Zhulenev 2025-12-18 02:13:37 -08:00
  • 2df2c4fac7 [XLA] Extend reshape-transpose chain removal to include bitcasts. Theotime Combes 2025-12-18 01:53:53 -08:00
  • 3d8a8b3367 PR #35353: [WIP ROCm] Fix flaky PersistedAutotuningTest.SingleOperationGetsAutotuned Aleksei Nurmukhametov 2025-12-18 01:43:15 -08:00
  • 5174b1f74c [XLA:GPU] Add Sort Fusion kind and corresponding FusionInterface. Adrian Kuegel 2025-12-18 01:22:36 -08:00
  • b79b6d8f75 Automated Code Change A. Unique TensorFlower 2025-12-18 01:14:34 -08:00
  • e560901dcd Reverts 6d3c0f702f A. Unique TensorFlower 2025-12-18 01:04:32 -08:00
  • 8c32a65652 compat: Update forward compatibility horizon to 2025-12-18 A. Unique TensorFlower 2025-12-18 01:03:50 -08:00
  • 4a142bc1a6 Update GraphDef version to 2445. A. Unique TensorFlower 2025-12-18 01:03:49 -08:00
  • cf9a56a83a Automated Code Change A. Unique TensorFlower 2025-12-18 01:01:41 -08:00
  • 5223bfde21 Automated Code Change A. Unique TensorFlower 2025-12-17 23:46:40 -08:00
  • d0ac32c16b Automated Code Change A. Unique TensorFlower 2025-12-17 23:14:39 -08:00
  • 08c7eea519 Automated Code Change A. Unique TensorFlower 2025-12-17 22:48:54 -08:00
  • 2bc359c7b1 Automated Code Change A. Unique TensorFlower 2025-12-17 21:30:13 -08:00
  • d5d3d8a868 Automated Code Change A. Unique TensorFlower 2025-12-17 21:22:29 -08:00
  • 1bbc0c679a Automated Code Change A. Unique TensorFlower 2025-12-17 21:16:01 -08:00
  • 066852ef91 Automated Code Change A. Unique TensorFlower 2025-12-17 21:09:14 -08:00
  • 702b294b41 Removal of tsl-specific integral types. A. Unique TensorFlower 2025-12-17 20:40:29 -08:00
  • be173da8ef Automated Code Change A. Unique TensorFlower 2025-12-17 20:39:56 -08:00
  • 8cfef0f4ff Automated Code Change A. Unique TensorFlower 2025-12-17 20:31:21 -08:00
  • 880c73f0a9 Automated Code Change A. Unique TensorFlower 2025-12-17 20:30:13 -08:00
  • f14f3b94e7 Automated Code Change A. Unique TensorFlower 2025-12-17 20:09:15 -08:00
  • ea54dd41da Automated Code Change A. Unique TensorFlower 2025-12-17 20:08:09 -08:00
  • 6a5e39020d Automated Code Change A. Unique TensorFlower 2025-12-17 20:01:20 -08:00
  • 0e07a3d2e3 Automated Code Change A. Unique TensorFlower 2025-12-17 20:01:16 -08:00
  • ad13e0db77 Automated Code Change A. Unique TensorFlower 2025-12-17 19:55:50 -08:00
  • af047e5a0d Automated Code Change A. Unique TensorFlower 2025-12-17 19:38:06 -08:00
  • b8d2866c35 Avoid redundant memset to clear the allocated backing store. Jeffrey A. Dean 2025-12-17 19:28:15 -08:00
  • ebdfe2dc37 Removed unused GetTaskState from coordination service. Michael Whittaker 2025-12-17 18:46:25 -08:00
  • 39d9238831 Update to use half data type in Cast kernel. Fengwu Yao 2025-12-17 17:58:54 -08:00
  • 0702b4623e Add Shape to ConvolutionThunk buffer_uses Maxim Ermilov 2025-12-17 17:52:10 -08:00
  • 18c6bdf2e7 Add mlir definition for PostProcessPrediction. A. Unique TensorFlower 2025-12-17 17:49:31 -08:00
  • ad2b228d6f [XLA:CPU] Add initial support of grouped convolutions with YNNPACK enabled. Volodymyr Kysenko 2025-12-17 17:13:58 -08:00
  • 07660093c8 Remove unused WaitForAllTasks from coordination service Michael Whittaker 2025-12-17 16:59:47 -08:00
  • 11ec4073f6 Handle cupti hardware trace correctly. Flip to true only once. A. Unique TensorFlower 2025-12-17 16:43:44 -08:00
  • 3015ca53fe Add proto serialization for AllReduceStartThunk Maxim Ermilov 2025-12-17 16:43:33 -08:00
  • 9f44d0753c Export sdy.replicated_to_unreduced to a manual computation. Zixuan Jiang 2025-12-17 15:42:44 -08:00
  • e1f8fd4ccb Add Google-specific signal handling. Michael Whittaker 2025-12-17 15:21:48 -08:00
  • 5c0a168ea1 Fix HloRunnerPjRt incorrectly not re-tupling results for replicated execution. Niklas Vangerow 2025-12-17 15:13:44 -08:00
  • 6d3c0f702f [stream_executor:cuda] Use Nccl/NvshmemMemoryAllocator to allocate collective memory A. Unique TensorFlower 2025-12-17 14:45:49 -08:00
  • 1262b408a2 [PjRt-IFRT] Internally track the output spec of ifrt::PjRtExecutable Hyeontaek Lim 2025-12-17 14:04:30 -08:00
  • a207484d9e Update to use half data type in split. Fengwu Yao 2025-12-17 13:58:35 -08:00
  • b525b848e7 Implement AbslStringify for strong int types in TSL Junwhan Ahn 2025-12-17 13:52:26 -08:00
  • 4350883de6 Add platform name to xla::ifrt::Device Ionel Gog 2025-12-17 13:44:21 -08:00
  • 7585d543b0 [PJRT] Change BuildPlanNodes, ChooseParallelizationStrategy, and the loop ordering code to look only at Loop objects, not other parts of the transpose plan. Peter Hawkins 2025-12-17 13:39:40 -08:00
  • b788805f73 Remove unused ReportError RPC from coordination service. Michael Whittaker 2025-12-17 13:36:40 -08:00
  • 26ebc05f4f Implement CommonPjRtLoadedExecutable::Execute, CommonPjRtLoadedExecutable::ExecutePortable and CommonPjRtLoadedExecutable::ExecuteSharded. Parker Schuh 2025-12-17 13:35:18 -08:00
  • a8f27858b8 Add Shape to DynamicSliceThunk buffer_uses Maxim Ermilov 2025-12-17 13:15:31 -08:00
  • 9f69553083 Run shardy/google/integrate_latest.sh. Zixuan Jiang 2025-12-17 13:00:38 -08:00
  • be344bbfb6 Breaking internal models in g3. Karlo Basioli 2025-12-17 12:55:55 -08:00
  • 336910b0b3 Add PoisonExecution to PjRtCApiDevice. Haibo Huang 2025-12-17 12:45:56 -08:00
  • fe02aab311 Supported int2 in xnnpack_delegate. Misha Gutman 2025-12-17 12:37:25 -08:00
  • 667249ad7a Update rules_ml_toolchain version to remove redundant fake_nvshmem_bootstrap_uid library from hermetic CUDA deps. Yulia Baturina 2025-12-17 12:31:50 -08:00
  • 5841fece45 Internal changes only. Fengwu Yao 2025-12-17 12:31:37 -08:00
  • a0aeb4fa6f Update rules_ml_toolchain version to remove redundant fake_nvshmem_bootstrap_uid library from hermetic CUDA deps. Yulia Baturina 2025-12-17 11:27:42 -08:00
  • 722bf3e739 Add ExecuteReplicatedWithExecutable to HloRunnerInterface. Niklas Vangerow 2025-12-17 11:06:48 -08:00
  • 413f4136c3 Update XNNPack version Quentin Khan 2025-12-17 11:04:18 -08:00
  • 356223b183 PR #34945: [ROCm] Add support for parametrized rocm hermetic dependency Alex 2025-12-17 10:43:06 -08:00
  • 24d56183bf PR #35339: Improve memory allocation error message Olli Lupton 2025-12-17 10:32:16 -08:00
  • 55afa175b0 [XLA:GPU] Add a utility to get GpuTargetConfig. Alexander Belyaev 2025-12-17 09:48:33 -08:00
  • e8cfd652f5 [XLA:GPU] Temporary revert statically registered collectives allocators since they are breaking OSS JAX tests due dynamic linking. A. Unique TensorFlower 2025-12-17 09:11:20 -08:00
  • 7375fe4a62 Automated Code Change Jie Luo 2025-12-17 09:04:40 -08:00
  • 95ef4892ba Remove unused gpu_types.h header and build target. Henning Becker 2025-12-17 08:58:26 -08:00
  • b22ae073cd Canonicalize convolutions with float inputs and non-float outputs by performing the convolution in F32 and casting the result back to the original output type. A. Unique TensorFlower 2025-12-17 08:23:47 -08:00
  • 3b2c89f5ed Always write a valid initial cache file when starting a cache build. Quentin Khan 2025-12-17 07:40:18 -08:00
  • b295c98d0a Automated Code Change A. Unique TensorFlower 2025-12-17 07:28:19 -08:00
  • a6d99fab4e Automated Code Change A. Unique TensorFlower 2025-12-17 07:26:00 -08:00
  • df6b63e83e Automated Code Change A. Unique TensorFlower 2025-12-17 07:24:37 -08:00
  • 114681a76c Automated Code Change A. Unique TensorFlower 2025-12-17 07:22:41 -08:00
  • f40e47ecab Automated Code Change A. Unique TensorFlower 2025-12-17 07:20:32 -08:00
  • 9c9429b376 PR #35113: Enqueue cross-host send after send buffer definition events are recorded, not complete Ashish Rao 2025-12-17 07:12:23 -08:00