JAXBench: Add back jaxbench profiling and block size tuning scripts#52
Open
charleshong3 wants to merge 2 commits into
Open
JAXBench: Add back jaxbench profiling and block size tuning scripts#52charleshong3 wants to merge 2 commits into
charleshong3 wants to merge 2 commits into
Conversation
Adds benchmark/profile_workload.py, a self-contained device-side timing
harness for JAXBench Pallas kernels. It JIT-compiles a workload variant
(baseline / optimized / custom), runs 5 warmup + 50 profiled iterations
under jax.profiler.trace(), parses the Perfetto JSON for per-iteration
on-device kernel times, and reports both device-profiler median_ms and a
wall_clock_median_ms cross-check, plus FLOPs/TFLOPs/utilization.
This is the authoritative device-side measurement method (vs host
wall-clock), suitable for apples-to-apples comparison against the
hand-tuned Pallas references. Only stdlib + jax/numpy deps.
Usage:
PJRT_DEVICE=TPU python3 benchmark/profile_workload.py benchmark/8p_GEMM baseline
PJRT_DEVICE=TPU python3 benchmark/profile_workload.py benchmark/8p_GEMM optimized
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds benchmark/tune_pallas.py, the companion grid-search tuner that
sweeps block-size configurations for each Pallas-optimized kernel
(e.g. GEMM block_shape/block_k, Paged pages_per_compute_block, Ragged
Paged Attention num_kv_pages_per_block/num_queries_per_block, Flash
block_q/block_k_major/block_k, Sparse block_q/block_kv), profiles each
config device-side, and reports the best. This is the tool that produced
the TUNED_PARAMS baked into the hand-tuned optimized.py references.
Self-contained (stdlib + jax/numpy), independent of profile_workload.py.
Usage:
PJRT_DEVICE=TPU python3 benchmark/tune_pallas.py [workload_name]
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA). View this failed invocation of the CLA check for more information. For the most up to date status, view the checks section at the bottom of the pull request. |
charleshong3
changed the title
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
These used to be in JAXBench but seem to have been removed during the migration to accelerator-agents. Not sure if we want them back.