[ROCm] upstream using rocprofiler-sdk (v3) for tracing AMD GPU events #29769
Conversation
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
Contributor
|
Hi @xla-rotation this PR is important to us, and it's amd-only code, could you review it please? thanks! |
beckerhe
reviewed
Aug 13, 2025
cj401-amd
force-pushed
the
ci_cj-rocprof-v3-roctracer-v1
branch
from
90fd457 to
c6080d1
Compare
cj401-amd
changed the title
beckerhe
approved these changes
Sep 10, 2025
akuegel
reviewed
Sep 11, 2025
cj401-amd
force-pushed
the
ci_cj-rocprof-v3-roctracer-v1
branch
from
c6080d1 to
8c23ab9
Compare
akuegel
approved these changes
Sep 18, 2025
dimitar-asenov
requested changes
Sep 19, 2025
Member
dimitar-asenov
left a comment
dimitar-asenov
left a comment
There was a problem hiding this comment.
Please #include in each source file only the files that are used in this file and all the files that are used in this file. E.g. rocm_tracer_utils.cc file at the moment includes only the corresponding header, but instead should include everything it uses directly. Similarly, rocm_tracer_utils.h should not include any files that it does not use directly. Please also do this for all the other files.
Another comment:
Do not use the <chrono> library. Use substitutes from absl instead.
cj401-amd
force-pushed
the
ci_cj-rocprof-v3-roctracer-v1
branch
from
63542a8 to
88fc01a
Compare
Contributor
Author
Hey @dimitar-asenov, please help have a look as we have addressed your comments now. btw, I wonder if it is possible to let me know what "Google internal checks FAILED" is? |
dimitar-asenov
requested changes
Sep 24, 2025
Member
Thank you for make the changes. I've flagged additional changes from our internal checks. Sorry for the back and forth, I hope this is the last round. |
copybara-service bot
pushed a commit
that referenced
this pull request
Sep 26, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- 99fe3f6 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 3f015cb by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 04e2b6b by cj401-amd <[email protected]>: update the macro -- 88fc01a by cj401-amd <[email protected]>: update header files -- e664bcb by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 e664bcb PiperOrigin-RevId: 808483340
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Sep 26, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- 99fe3f655750b5deba0b5dd39d0af953b6d5f189 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 3f015cbba8772290e227dbc00f0fd3c010c946cc by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 04e2b6b1e92105933dfc1182d5fc8a15b7fdef87 by cj401-amd <[email protected]>: update the macro -- 88fc01aa569a86dacb9332403a36eec2a7f9a6db by cj401-amd <[email protected]>: update header files -- e664bcb687bfb377b2219e742778a351df523cda by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 e664bcb687bfb377b2219e742778a351df523cda PiperOrigin-RevId: 808483340
copybara-service bot
pushed a commit
that referenced
this pull request
Sep 26, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- 99fe3f6 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 3f015cb by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 04e2b6b by cj401-amd <[email protected]>: update the macro -- 88fc01a by cj401-amd <[email protected]>: update header files -- e664bcb by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 e664bcb PiperOrigin-RevId: 808483340
copybara-service bot
pushed a commit
that referenced
this pull request
Sep 26, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- 99fe3f6 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 3f015cb by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 04e2b6b by cj401-amd <[email protected]>: update the macro -- 88fc01a by cj401-amd <[email protected]>: update header files -- e664bcb by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 e664bcb PiperOrigin-RevId: 808483340
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Sep 26, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- 99fe3f655750b5deba0b5dd39d0af953b6d5f189 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 3f015cbba8772290e227dbc00f0fd3c010c946cc by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 04e2b6b1e92105933dfc1182d5fc8a15b7fdef87 by cj401-amd <[email protected]>: update the macro -- 88fc01aa569a86dacb9332403a36eec2a7f9a6db by cj401-amd <[email protected]>: update header files -- e664bcb687bfb377b2219e742778a351df523cda by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 e664bcb687bfb377b2219e742778a351df523cda PiperOrigin-RevId: 808483340
Contributor
Author
Hi @dimitar-asenov, I wonder if it is possible to let me know the internal check for this PR, so I can adapt the corresponding parts. Many thanks. |
zahiqbal
pushed a commit
to ROCm/xla
that referenced
this pull request
Oct 1, 2025
upstream PR: openxla/pull/29769 Squash following commits.. Update rocprofiler-sdk (v3) along with roctracer (v1) for rocm-jaxlib-v0.6.0 (#302) * update for integration of rocprofiler-sdk (along with roctracer as a backup based on bazel_options from CLI) (cherry picked from commit 7775dd0) use VLOG(2) to replace LOG(INFO), so PGLE has no verbose info (#357) (cherry picked from commit 5950125) update with kernel details for rocm-7.x (#364) * update with kernel details for rocm-7.x (cherry picked from commit 5597c0d) update to remove previously hard-coded rocprofiler-sdk path (#369) * update to remove previously hard-coded rocprofiler-sdk path and add skip_rocprofiler_sdk to avoid loading `rocprofiler-sdk` (cherry picked from commit ff74b5f)
zahiqbal
pushed a commit
to ROCm/xla
that referenced
this pull request
Oct 1, 2025
upstream PR: openxla/pull/29769 Squash following commits.. Update rocprofiler-sdk (v3) along with roctracer (v1) for rocm-jaxlib-v0.6.0 (#302) * update for integration of rocprofiler-sdk (along with roctracer as a backup based on bazel_options from CLI) (cherry picked from commit 7775dd0) use VLOG(2) to replace LOG(INFO), so PGLE has no verbose info (#357) (cherry picked from commit 5950125) update with kernel details for rocm-7.x (#364) * update with kernel details for rocm-7.x (cherry picked from commit 5597c0d) update to remove previously hard-coded rocprofiler-sdk path (#369) * update to remove previously hard-coded rocprofiler-sdk path and add skip_rocprofiler_sdk to avoid loading `rocprofiler-sdk` (cherry picked from commit ff74b5f)
zahiqbal
pushed a commit
to ROCm/xla
that referenced
this pull request
Oct 2, 2025
upstream PR: openxla/pull/29769 Squash following commits.. Update rocprofiler-sdk (v3) along with roctracer (v1) for rocm-jaxlib-v0.6.0 (#302) * update for integration of rocprofiler-sdk (along with roctracer as a backup based on bazel_options from CLI) (cherry picked from commit 7775dd0) use VLOG(2) to replace LOG(INFO), so PGLE has no verbose info (#357) (cherry picked from commit 5950125) update with kernel details for rocm-7.x (#364) * update with kernel details for rocm-7.x (cherry picked from commit 5597c0d) update to remove previously hard-coded rocprofiler-sdk path (#369) * update to remove previously hard-coded rocprofiler-sdk path and add skip_rocprofiler_sdk to avoid loading `rocprofiler-sdk` (cherry picked from commit ff74b5f)
1 task
zahiqbal
pushed a commit
to ROCm/xla
that referenced
this pull request
Oct 3, 2025
upstream PR: openxla/pull/29769 Squash following commits.. Update rocprofiler-sdk (v3) along with roctracer (v1) for rocm-jaxlib-v0.6.0 (#302) * update for integration of rocprofiler-sdk (along with roctracer as a backup based on bazel_options from CLI) (cherry picked from commit 7775dd0) use VLOG(2) to replace LOG(INFO), so PGLE has no verbose info (#357) (cherry picked from commit 5950125) update with kernel details for rocm-7.x (#364) * update with kernel details for rocm-7.x (cherry picked from commit 5597c0d) update to remove previously hard-coded rocprofiler-sdk path (#369) * update to remove previously hard-coded rocprofiler-sdk path and add skip_rocprofiler_sdk to avoid loading `rocprofiler-sdk` (cherry picked from commit ff74b5f)
hsharsha
pushed a commit
to ROCm/xla
that referenced
this pull request
Oct 6, 2025
* rocprof-sdk addition, upstream PR: openxla/pull/29769 Squash following commits.. Update rocprofiler-sdk (v3) along with roctracer (v1) for rocm-jaxlib-v0.6.0 (#302) * update for integration of rocprofiler-sdk (along with roctracer as a backup based on bazel_options from CLI) (cherry picked from commit 7775dd0) use VLOG(2) to replace LOG(INFO), so PGLE has no verbose info (#357) (cherry picked from commit 5950125) update with kernel details for rocm-7.x (#364) * update with kernel details for rocm-7.x (cherry picked from commit 5597c0d) update to remove previously hard-coded rocprofiler-sdk path (#369) * update to remove previously hard-coded rocprofiler-sdk path and add skip_rocprofiler_sdk to avoid loading `rocprofiler-sdk` (cherry picked from commit ff74b5f) * fixed buffer comparator test --------- Co-authored-by: Chunyu Jin <[email protected]>
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
beckerhe
reviewed
Nov 5, 2025
| Returns: | ||
| Map of library names to structs of filename and path | ||
| """ | ||
| repo_path = str(repository_ctx.path(rocm_config.rocm_toolkit_path)) |
Member
There was a problem hiding this comment.
This change seems to lead to a build failure, see https://github.com/openxla/xla/actions/runs/19111720943/job/54610280919
But I'm also not sure how this ever worked given that there is a type in line 335 (rep_path instead of repo_path).
Contributor
Author
There was a problem hiding this comment.
@beckerhe, not it's been fixed as locally bazel build //xla/backends/profiler/gpu:rocm_tracer --config=rocm_ci_hermetic
DEBUG: /work/xla/third_party/py/python_repo.bzl:89:10:
=============================
Hermetic Python configuration:
Version: "3.11"
Kind: ""
Interpreter: "default" (provided by rules_python)
Requirements_lock label: "@//:requirements_lock_3_11.txt"
=====================================
Computing main repo mapping:
Computing main repo mapping:
Computing main repo mapping:
Computing main repo mapping:
Computing main repo mapping:
Loading:
Loading: 3 packages loaded
DEBUG: /root/.cache/bazel/_bazel_root/ea1efa0977f8828bf242d5b6a382af7f/external/rules_ml_toolchain/third_party/gpus/cuda/hermetic/cuda_configure.bzl:97:14: Cannot find clang, either correct your path, or set the CLANG_CUDA_COMPILER_PATH or CC environment variables
Analyzing: target //xla/backends/profiler/gpu:rocm_tracer (4 packages loaded, 0 targets configured)
Analyzing: target //xla/backends/profiler/gpu:rocm_tracer (4 packages loaded, 0 targets configured)
Analyzing: target //xla/backends/profiler/gpu:rocm_tracer (52 packages loaded, 11 targets configured)
Analyzing: target //xla/backends/profiler/gpu:rocm_tracer (52 packages loaded, 11 targets configured)
Analyzing: target //xla/backends/profiler/gpu:rocm_tracer (118 packages loaded, 2381 targets configured)
Analyzing: target //xla/backends/profiler/gpu:rocm_tracer (191 packages loaded, 20354 targets configured)
INFO: Analyzed target //xla/backends/profiler/gpu:rocm_tracer (191 packages loaded, 20450 targets configured).
[2,494 / 3,466] [Prepa] Compiling absl/log/internal/structured_proto.cc ... (128 actions, 3 running)
[2,544 / 3,466] Compiling absl/log/internal/vlog_config.cc; 0s local ... (126 actions, 125 running)
[2,639 / 3,466] Compiling tsl/platform/stringprintf.cc; 1s local ... (128 actions, 127 running)
[2,685 / 3,466] Compiling tsl/platform/cpu_info.cc; 2s local ... (129 actions, 127 running)
[2,771 / 3,466] Compiling src/google/protobuf/compiler/main.cc [for tool]; 3s local ... (128 actions, 125 running)
[2,871 / 3,466] Compiling src/google/protobuf/descriptor.pb.cc; 4s local ... (128 actions, 125 running)
[2,942 / 3,466] Compiling src/google/protobuf/descriptor.pb.cc; 5s local ... (128 actions, 127 running)
[2,972 / 3,466] Compiling src/google/protobuf/descriptor.cc; 6s local ... (128 actions, 127 running)
[2,984 / 3,466] Compiling src/google/protobuf/descriptor.cc; 7s local ... (129 actions, 128 running)
[2,995 / 3,466] Compiling src/google/protobuf/descriptor.cc; 8s local ... (128 actions, 127 running)
[3,029 / 3,466] Compiling src/google/protobuf/descriptor.cc; 9s local ... (129 actions, 128 running)
[3,086 / 3,466] Compiling src/google/protobuf/descriptor.cc; 10s local ... (128 actions, 127 running)
[3,142 / 3,466] Compiling src/google/protobuf/descriptor.cc; 11s local ... (129 actions, 127 running)
[3,319 / 3,466] Compiling src/google/protobuf/descriptor.cc; 13s local ... (21 actions running)
[3,329 / 3,466] Compiling src/google/protobuf/descriptor.cc; 14s local ... (11 actions running)
[3,335 / 3,466] Compiling src/google/protobuf/descriptor.cc; 15s local ... (5 actions running)
[3,337 / 3,466] Compiling src/google/protobuf/descriptor.cc; 16s local ... (3 actions running)
[3,339 / 3,466] Compiling src/google/protobuf/descriptor.cc [for tool]; 9s local
[3,340 / 3,466] [Prepa] Linking external/com_google_protobuf/upb_generator/c/protoc-gen-upb_stage0 [for tool] ... (3 actions, 0 running)
[3,367 / 3,466] Compiling upb_generator/minitable/fasttable.cc [for tool]; 1s local ... (5 actions running)
[3,371 / 3,466] Compiling upb_generator/minitable/generator.cc [for tool]; 2s local
[3,410 / 3,466] Compiling src/google/protobuf/util/field_mask_util.cc; 1s local ... (53 actions running)
[3,430 / 3,466] Compiling xla/tsl/platform/env.cc; 2s local ... (33 actions running)
[3,447 / 3,466] Compiling src/google/protobuf/json/internal/parser.cc; 3s local ... (18 actions, 17 running)
[3,465 / 3,466] Compiling xla/stream_executor/rocm/rocm_executor.cc; 4s local
INFO: Found 1 target...
Target //xla/backends/profiler/gpu:rocm_tracer up-to-date:
bazel-bin/xla/backends/profiler/gpu/librocm_tracer.a
bazel-bin/xla/backends/profiler/gpu/librocm_tracer.pic.a
bazel-bin/xla/backends/profiler/gpu/librocm_tracer.so
INFO: Elapsed time: 149.625s, Critical Path: 22.86s
INFO: 3466 processes: 2500 internal, 966 local.
INFO: Build completed successfully, 3466 total actions
Contributor
There was a problem hiding this comment.
This merge broke our hermetic build
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally -- 757f064 by cj401-amd <[email protected]>: update for fixing the typo Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 757f064 PiperOrigin-RevId: 828424223
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally -- 757f064e88b4047e6bcff32513074ce5f57522a0 by cj401-amd <[email protected]>: update for fixing the typo Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 757f064e88b4047e6bcff32513074ce5f57522a0 PiperOrigin-RevId: 828424223
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally -- 757f064 by cj401-amd <[email protected]>: update for fixing the typo Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 757f064 PiperOrigin-RevId: 828424223
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally -- 757f064e88b4047e6bcff32513074ce5f57522a0 by cj401-amd <[email protected]>: update for fixing the typo Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 757f064e88b4047e6bcff32513074ce5f57522a0 PiperOrigin-RevId: 828424223
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally -- 757f064 by cj401-amd <[email protected]>: update for fixing the typo Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 757f064 PiperOrigin-RevId: 828424223
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 5, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 6, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 6, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 6, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 6, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249dbf7978425bf0afe0c12f952d557f2943e PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
that referenced
this pull request
Nov 6, 2025
… GPU events Imported from GitHub PR #29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab06442 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5c by cj401-amd <[email protected]>: update the macro -- cb13f7e by cj401-amd <[email protected]>: update header files -- 15ec380 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f3 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249d by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 FUTURE_COPYBARA_INTEGRATE_REVIEW=#29769 from ROCm:ci_cj-rocprof-v3-roctracer-v1 ac2249d PiperOrigin-RevId: 828431731
copybara-service bot
pushed a commit
to tensorflow/tensorflow
that referenced
this pull request
Nov 6, 2025
… GPU events Imported from GitHub PR openxla/xla#29769 @xla-rotation, would you please kindly help review this PR? We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti. This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs. - Integrate rocprofiler-sdk (v3) to XLA for improved profiling of GPU events, support both time-based and step-based profiling, - Add unit tests for rocm_collector and rocm_tracer Copybara import of the project: -- ab064425550acf959ec7b6d404b294b6bc8a34f3 by cj401-amd <[email protected]>: update with refactoring based on the PR comments -- 8f8ad5a5d77e2930a7489815baf3bf2bfbac7558 by cj401-amd <[email protected]>: update with VLOG(2) and fix a bug, new line at the EoF -- 6a85e5cfc4b1696b47df94b7e1082567e3cdac8a by cj401-amd <[email protected]>: update the macro -- cb13f7e5cce2e9e64efa224ca35cfecf5fcdf5ba by cj401-amd <[email protected]>: update header files -- 15ec380fb3a6e7e053d3d466496496451a2121e8 by cj401-amd <[email protected]>: update a: addressing comments, b: add kernel details, c: update rocprofiler-sdk path -- 3ee71f30300c482de2730c39bb5568d3b5f58509 by cj401-amd <[email protected]>: add rocprofiler-sdk debs for hermetic build -- ac2249dbf7978425bf0afe0c12f952d557f2943e by cj401-amd <[email protected]>: update with a successful hermetic build locally Merging this change closes #29769 PiperOrigin-RevId: 828829024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 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.
@xla-rotation, would you please kindly help review this PR?
We are phasing out development and support for roctracer/rocprofiler/rocprof/rocprofv2 in favor of rocprofiler-sdk (v3) in upcoming ROCm releases. rocprofielr-sdk (v3) also moves away from cupti.
This PR integrates rocprofiler-sdk (v3) into XLA for profiling GPU events on AMD GPUs.