[pull] master from tensorflow:master#265
Merged
pull[bot] merged 15 commits intoIrvanFza:masterfrom Jul 30, 2025
Merged
Conversation
… the constraints. The original issue manifested in the form of CUDA_ERROR_ILLEGAL_INSTRUCTION on microbenchmarks that used TMA. The constraint is undocumented in Nvidia's docs, and was informed via chat when the issue was raised. PiperOrigin-RevId: 788827881
This is a somewhat tricky case as scatter reuses buffers for operands for the corresponding output. So if we have the same repeated operand, we should instead transform that into having copies of the operand for the repetitions. PiperOrigin-RevId: 788836845
We were calling ComposePermutations where a permutation was applied to a non-permutation. This should be done with the method Permute() instead. This is a non-functional change in opt builds, but before this change, a DCHECK was triggered. PiperOrigin-RevId: 788860614
PiperOrigin-RevId: 788869943
…has passed. PiperOrigin-RevId: 788871192
PiperOrigin-RevId: 788871838
…if no layout has been set. Sometimes the `HloModules` have the `entry_computation_layout` field set with no layouts specified, and so we shouldn't exit early returning the current field. As we may need to tuplify it if `useTupleArgs` is true PiperOrigin-RevId: 788875871
PiperOrigin-RevId: 788876921
Imported from GitHub PR openxla/xla#29342 The XLA Linux x86 GPU ONEAPI pre-submit is failing because certain flags are being overridden, which is blocking the hermetic build from running successfully. Copybara import of the project: -- 9e4190e4c627e853ddd9dbcde936f733101c1211 by mraunak <mayank.kumar.raunak@intel.com>: Create ci_build_xla.sh Build with Clang -- bdc183da860a31f1be2f6dd3858a8ac44a53f9d2 by mraunak <mayank.kumar.raunak@intel.com>: Update configure.py -- 1a874cf69fd84d69b26ac4e37843887ac8291bb8 by mraunak <mayank.kumar.raunak@intel.com>: Add Clang as host compiler for oneAPI -- 342b1d67645f6fedc9ad48ea699fe558e273530b by mraunak <mayank.kumar.raunak@intel.com>: Update sycl_configure.bzl -- 2f38e0d154977ecac50bdb220060789a0f835c22 by mraunak <mayank.kumar.raunak@intel.com>: Update BUILD.sycl.tpl -- 7df8e577fea5861b2b28cc18906b289b4e9ac5df by mraunak <mayank.kumar.raunak@intel.com>: Update ci_build_xla.sh -- bef3a5ca6faf1378636de287996a70ce8177cb95 by mraunak <mayank.kumar.raunak@intel.com>: Update sycl_configure.bzl -- a960ff7855f5dcefebd0913d7b7712e579028cac by mraunak <mayank.kumar.raunak@intel.com>: Update tensorflow.bazelrc -- 2d60209a0333204c5abb907501abff4d0af21271 by mraunak <mayank.kumar.raunak@intel.com>: Update ci_test_xla.sh -- 8a5db012f9f1e664711c3fbb90c7215f21effd2d by mraunak <mayank.kumar.raunak@intel.com>: Address the comments -- 735f4591654ee377f808001959cd566b4b6a0544 by mraunak <mayank.kumar.raunak@intel.com>: Address the comments -- e7c412d19b25804c4fa4fb0b65329146c983a6e3 by mraunak <mayank.kumar.raunak@intel.com>: Address the comment -- b456c8f8dca013bb730a733b1f988646a0a744bb by mraunak <mayank.kumar.raunak@intel.com>: Update tensorflow.bazelrc -- cedc8586494f789e5ec92aef30e64e866a6b022d by mraunak <mayank.kumar.raunak@intel.com>: Apply patch to PR -- d2dc3846a7fdd55f35fc66f314ad56265e963c59 by mraunak <mayank.kumar.raunak@intel.com>: fix the test failure -- 33a3ca34306b6047881f499aa76fc69b4f5003a6 by mraunak <mayank.kumar.raunak@intel.com>: Resolve merge conflicts -- 329b71e7c72e1df001f9d104c25f44032739cc02 by mraunak <mayank.kumar.raunak@intel.com>: Update sycl_configure.bzl -- 253cdd753760e8718973b4493f6aafc07205bb2d by mraunak <mayank.kumar.raunak@intel.com>: Fix the pre-submit failure -- 326fb3a4bbaf6e62caaf8f2135405540b0901739 by mraunak <mayank.kumar.raunak@intel.com>: Update configure.py -- fea0523b3afc170ba7a9a7a7b72d7327c5aa2637 by mraunak <mayank.kumar.raunak@intel.com>: Update sycl_configure.bzl -- e841d90370dc2c45187df8a8965dde5d137a8455 by mraunak <mayank.kumar.raunak@intel.com>: Update configure.py -- 9d6e9d7512c4f6bd803d4ee0297dbb2306179c67 by mraunak <mayank.kumar.raunak@intel.com>: Fix the indentation issue Merging this change closes #29342 PiperOrigin-RevId: 788890366
Imported from GitHub PR openxla/xla#29332 A bug was reported in NVIDIA that GpuAotCompilationTest was failing, specifically these two ``` [ FAILED ] GpuAotCompilationTest.ExportAndLoadExecutable [ FAILED ] GpuAotCompilationTest.AotCompilationWithoutGpuDevice ``` with the error message ``` NOT_FOUND: Could not find registered platform with name: "CUDA". Available platform names are: Interpreter ``` The cause is the cuda platform was not properly registered. Modified the BUILD file to link the cuda platform to make sure it is registered before executing the test. Copybara import of the project: -- a245bd2958e9794d19ecd87306eb298e62015f75 by Chenhao Jiang <chenhaoj@nvidia.com>: Fix the gpu_aot_compilation_test Merging this change closes #29332 PiperOrigin-RevId: 788893566
PiperOrigin-RevId: 788895133
…ise int4 A100 tests are failing. In the cl we pass device description to the rewriter and check what exact device we have. PiperOrigin-RevId: 788908460
PiperOrigin-RevId: 788911526
PiperOrigin-RevId: 788914574
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
11 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.3)
Can you help keep this open source service alive? 💖 Please sponsor : )