[ROCM] Add nanoo fp8 support in type traits #17890
Closed
Conversation
This file contains 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
|
FYI: For the context on NANOO FP8, please refer to this comment: |
|
Hi @majnemer could you take a look at this PR? |
|
Hi @xla-rotation could you check this PR please? |
ezhulenev
approved these changes
Oct 11, 2024
copybara-service bot
pushed a commit
that referenced
this pull request
Oct 11, 2024
Imported from GitHub PR #17890 We are trying to add NANOO FP8 support in TensorFlow. To achieve this, a small modification in the `type_traits.h` file on the XLA side is required, where the NANOO FP8 should also be classified as a simple type. This change will be utilized in tensor.cc within TensorFlow. You can find the relevant code here: [TensorFlow tensor.cc](https://github.com/tensorflow/tensorflow/blob/ba7f93b8dc32b64cc0bb5ddb44e75bd344ae67d0/tensorflow/core/framework/tensor.cc#L175). Copybara import of the project: -- 1c27aeb by scxfjiang <xuefei.jiang@amd.com>: add nanoo fp8 support in type traits Merging this change closes #17890 Reverts changelist 673529458 FUTURE_COPYBARA_INTEGRATE_REVIEW=#17890 from ROCm:ci_nanoo_fp8_in_type_traits 1c27aeb PiperOrigin-RevId: 684839620
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Oct 11, 2024
Imported from GitHub PR openxla/xla#17890 We are trying to add NANOO FP8 support in TensorFlow. To achieve this, a small modification in the `type_traits.h` file on the XLA side is required, where the NANOO FP8 should also be classified as a simple type. This change will be utilized in tensor.cc within TensorFlow. You can find the relevant code here: [TensorFlow tensor.cc](https://github.com/tensorflow/tensorflow/blob/ba7f93b8dc32b64cc0bb5ddb44e75bd344ae67d0/tensorflow/core/framework/tensor.cc#L175). Copybara import of the project: -- 1c27aebbc206559861ee8d5e5f44522a2a44858b by scxfjiang <xuefei.jiang@amd.com>: add nanoo fp8 support in type traits Merging this change closes #17890 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#17890 from ROCm:ci_nanoo_fp8_in_type_traits 1c27aebbc206559861ee8d5e5f44522a2a44858b PiperOrigin-RevId: 684839620
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Oct 11, 2024
Imported from GitHub PR openxla/xla#17890 We are trying to add NANOO FP8 support in TensorFlow. To achieve this, a small modification in the `type_traits.h` file on the XLA side is required, where the NANOO FP8 should also be classified as a simple type. This change will be utilized in tensor.cc within TensorFlow. You can find the relevant code here: [TensorFlow tensor.cc](https://github.com/tensorflow/tensorflow/blob/ba7f93b8dc32b64cc0bb5ddb44e75bd344ae67d0/tensorflow/core/framework/tensor.cc#L175). Copybara import of the project: -- 1c27aebbc206559861ee8d5e5f44522a2a44858b by scxfjiang <xuefei.jiang@amd.com>: add nanoo fp8 support in type traits Merging this change closes #17890 PiperOrigin-RevId: 684917294
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Oct 11, 2024
…placements in memory space assignment. Both features are enabled by default. FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#17890 from ROCm:ci_nanoo_fp8_in_type_traits 1c27aebbc206559861ee8d5e5f44522a2a44858b PiperOrigin-RevId: 680765291
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Oct 11, 2024
FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#17890 from ROCm:ci_nanoo_fp8_in_type_traits 1c27aebbc206559861ee8d5e5f44522a2a44858b PiperOrigin-RevId: 684189732
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Oct 11, 2024
- Moves the 3 overloads into the subclasses - Moves the corresponding GpuDriver functions - Adds basic unit tests FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#17890 from ROCm:ci_nanoo_fp8_in_type_traits 1c27aebbc206559861ee8d5e5f44522a2a44858b PiperOrigin-RevId: 684699241
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
We are trying to add NANOO FP8 support in TensorFlow. To achieve this, a small modification in the
type_traits.hfile on the XLA side is required, where the NANOO FP8 should also be classified as a simple type. This change will be utilized in tensor.cc within TensorFlow. You can find the relevant code here: TensorFlow tensor.cc.