OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

Update Kineto submodule with fix to test_basic_chrome_trace (#128333)

Browse Source 
Summary: We've updated the sort_index in Kineto chrome traces to support device ids up to 16 devices. This should make chrome trace rows be ordered in the same way as CUDA. We need to update the unit test as well.

Test Plan:
Ran locally the changing test:
```
$ buck2 test 'fbcode//mode/opt' fbcode//caffe2/test:test_profiler_cuda -- --exact 'caffe2/test:test_profiler_cuda - test_basic_chrome_trace (profiler.test_profiler.TestProfiler)'
File changed: fbcode//caffe2/third_party/kineto.submodule.txt
Buck UI: https://www.internalfb.com/buck2/f4fd1e9a-99f1-4422-aeed-b54903c64146
Test UI: https://www.internalfb.com/intern/testinfra/testrun/16888498639845776
Network: Up: 5.4KiB  Down: 8.6KiB  (reSessionID-0329120e-7fa2-4bc0-b539-7e58058f8fce)
Jobs completed: 6. Time elapsed: 1:01.2s.
Tests finished: Pass 2. Fail 0. Fatal 0. Skip 0. Build failure 0
```

Differential Revision: D58362964

Pulled By: aaronenyeshi

Pull Request resolved: https://github.com/pytorch/pytorch/pull/128333
Approved by: https://github.com/Skylion007
This commit is contained in:
Aaron Enye Shi 2024-06-10 18:12:32 +00:00 committed by PyTorch MergeBot
parent 83941482f7
commit 136bdb96cb
2 changed files with 5 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
test/profiler/test_profiler.py
View File

@ -1723,9 +1723,12 @@ assert KinetoStepTracker.current_step() == initial_step + 2 * niters
gpu_value = traceEvent.get("args", {}).get("labels", None)
if gpu_value and "GPU" in gpu_value:
gpu_dict[gpu_value] += 1
# Max PID offset is 5M, based from pytorch/kineto include header:
# https://github.com/pytorch/kineto/blob/8681ff11e1fa54da39023076c5c43eddd87b7a8a/libkineto/include/output_base.h#L35
kExceedMaxPid = 5000000
self.assertTrue(
traceEvents[i + 1]["args"]["sort_index"]
== 0x1000000 + int(gpu_value.split()[1])
== kExceedMaxPid + int(gpu_value.split()[1])
)
# TODO add checking gpu count if cpuOnly_ is true or not

2
third_party/kineto vendored

@ -1 +1 @@
Subproject commit be1317644c68b4bfc4646024a6b221066e430031
Subproject commit 8681ff11e1fa54da39023076c5c43eddd87b7a8a