-
Notifications
You must be signed in to change notification settings - Fork 424
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[GPU] Expose caching DebugOptions to python #15636
Conversation
a75829b
to
207bcc6
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is it possible to test those?
Thanks for reviewing, added test to |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good, please address the test comment though
@@ -3137,6 +3137,9 @@ def testCompileOptionsSerialization(self): | |||
executable_build_options.num_partitions = 2 | |||
executable_build_options.debug_options.xla_cpu_enable_fast_math = True | |||
executable_build_options.debug_options.xla_test_all_input_layouts = True | |||
executable_build_options.debug_options.xla_gpu_kernel_cache_file = "/foo/bar" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think you actually need to test these flags by adding them to L3157
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for the review! Fixed this.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sweet thanks!
Uh @cheshire should you get this in or should I? Not sure how to do xla PRs but I can try...
Imported from GitHub PR openxla/xla#15636 Adds python bindings for `xla_gpu_kernel_cache_file`, `xla_gpu_enable_llvm_module_compilation_parallelism` and `xla_gpu_per_fusion_autotune_cache_dir`. We would like to add some convenience features to JAX which will enable all caches with one flag/option (will open PR for that soon). This change is necessary for that. Copybara import of the project: -- 87dbb3d3abe5969d1509d965071c5c7dbdcb46d0 by Trevor Morris <tmorris@nvidia.com>: Expose caching DebugOptions to python -- 3bb41fca800bc12d00870e7b8cbde3641de9beab by Trevor Morris <tmorris@nvidia.com>: Add test -- f0e04233522d8740c83b9b19138b76304b242604 by Trevor Morris <tmorris@nvidia.com>: Fix tests Merging this change closes #15636 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#15636 from trevor-m:cache-python f0e04233522d8740c83b9b19138b76304b242604 PiperOrigin-RevId: 665300777
Imported from GitHub PR openxla/xla#15636 Adds python bindings for `xla_gpu_kernel_cache_file`, `xla_gpu_enable_llvm_module_compilation_parallelism` and `xla_gpu_per_fusion_autotune_cache_dir`. We would like to add some convenience features to JAX which will enable all caches with one flag/option (will open PR for that soon). This change is necessary for that. Copybara import of the project: -- 87dbb3d3abe5969d1509d965071c5c7dbdcb46d0 by Trevor Morris <tmorris@nvidia.com>: Expose caching DebugOptions to python -- 3bb41fca800bc12d00870e7b8cbde3641de9beab by Trevor Morris <tmorris@nvidia.com>: Add test -- f0e04233522d8740c83b9b19138b76304b242604 by Trevor Morris <tmorris@nvidia.com>: Fix tests Merging this change closes #15636 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#15636 from trevor-m:cache-python f0e04233522d8740c83b9b19138b76304b242604 PiperOrigin-RevId: 665300777
Imported from GitHub PR openxla/xla#15636 Adds python bindings for `xla_gpu_kernel_cache_file`, `xla_gpu_enable_llvm_module_compilation_parallelism` and `xla_gpu_per_fusion_autotune_cache_dir`. We would like to add some convenience features to JAX which will enable all caches with one flag/option (will open PR for that soon). This change is necessary for that. Copybara import of the project: -- 87dbb3d3abe5969d1509d965071c5c7dbdcb46d0 by Trevor Morris <tmorris@nvidia.com>: Expose caching DebugOptions to python -- 3bb41fca800bc12d00870e7b8cbde3641de9beab by Trevor Morris <tmorris@nvidia.com>: Add test -- f0e04233522d8740c83b9b19138b76304b242604 by Trevor Morris <tmorris@nvidia.com>: Fix tests Merging this change closes #15636 PiperOrigin-RevId: 665475870
Adds python bindings for
xla_gpu_kernel_cache_file
,xla_gpu_enable_llvm_module_compilation_parallelism
andxla_gpu_per_fusion_autotune_cache_dir
.We would like to add some convenience features to JAX which will enable all caches with one flag/option (will open PR for that soon). This change is necessary for that.