Added CPU runtime benchmarking and expanded BenchmarkResult to store CPU runtimes #3206
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
…ark_core` (pytorch#3191) Summary: Removed redundant logic in `benchmark` and `benchmark_func` by moving shared timing, memory, and profiling code into a new `_run_benchmark_core` function. Ensured the backward-compatibility. Differential Revision: D78290979
…CPU runtimes Summary: * Added feature to becnhmark CPU runtimes alongside with GPU measurements. * Expanded `BenchmarkResult` class to store both device measurements * Adapted files that are importing `BenchmarkResult` to ensure compatibility Now we can compare CPU and GPU runtimes without running and analyzing PyTorch Profiler. BenchmarkResults can help users to detect if the module/function/operator is CPU-bounded or GPU-bounded. Example metrics of FBGEMM operators: | Operator | CPU Runtime | GPU Runtime | GPU Memory | |---------------------------------------------|-------------|-------------|------------| | **[fallback] pytorch generic** | 5.41 ms | 2.66 ms | 1.01 GB | | **[Prod] KeyedTensor.regroup_dup** | 2.13 ms | 2.48 ms | 1.01 GB | | **[Module] KTRegroupAsDict_dup** | 0.14 ms | 0.75 ms | 1.01 GB | | **[2 Ops] permute_multi_embs_dup** | 0.88 ms | 1.44 ms | 1.01 GB | | **[1 Op] KT_regroup_dup** | 0.99 ms | 1.54 ms | 1.01 GB | We can see that `[fallback] pytorch generic` is CPU-bounded Differential Revision: D78503319
meta-cla
bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D78503319 |
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Outdated suggestions cannot be applied.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.
Summary:
BenchmarkResultclass to store both device measurementsBenchmarkResultto ensure compatibilityNow we can compare CPU and GPU runtimes without running and analyzing PyTorch Profiler. BenchmarkResults can help users to detect if the module/function/operator is CPU-bounded or GPU-bounded.
Example metrics of FBGEMM operators:
We can see that
[fallback] pytorch genericis CPU-boundedDifferential Revision: D78503319