This repository has been archived by the owner on Oct 11, 2024. It is now read-only.
Upstream sync 2024 06 16 #307
Merged
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
…st configurations (vllm-project#5253)
…ation test configurations" (vllm-project#5463)
…t fail for GPU tests (vllm-project#5464) Signed-off-by: kevin <[email protected]>
Inspired by vllm-project#5146, this PR improves FP8 quantize kernel by vectorizing data transfer to better utilize memory bandwidth. Microbenchmark shows that this improved kernel can achieve 1.0x-1.5x speedup (especially when hidden size is large). In details, we applied 3 optimizations: - Use inverted scale so that most divisions are changed to multiplications. - Unroll the loop by 4 times to improve ILP. - Use vectorized 4 to transfer data between HBM and SRAM.
…#4990) Signed-off-by: Travis Johnson <[email protected]> Co-authored-by: Sanger Steel <[email protected]> Co-authored-by: Roger Wang <[email protected]>
…-project#5293) [Core][Distributed] add coordinator to reduce code duplication in tp and pp (vllm-project#5293)
Signed-off-by: kevin <[email protected]>
…should i… (vllm-project#5303) Signed-off-by: Wang, Yi A <[email protected]> Co-authored-by: Roger Wang <[email protected]>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-neuralmagic@users.noreply.github.com>
…n test configurations (vllm-project#5466)
…#5497) Tune Qwen2-57B-A14B configs based on vllm-project#4921 Throughput Performance command: python benchmarks/benchmark_throughput.py --model=Qwen/Qwen2-57B-A14B-Instruct --input-len 1000 --output-len 50 -tp 2 A100 GPU benchmark no config w/ PR tp=2 10.53 requests/s, 11058.17 tokens/s 12.47 requests/s, 13088.57 tokens/s tp=4 17.77 requests/s, 18662.95 tokens/s 20.20 requests/s, 21212.32 tokens/s
…llm-project#4971) Co-authored-by: Jianan Gu <[email protected]>
Co-authored-by: Roger Wang <[email protected]>
Co-authored-by: Michael Goin <[email protected]> Co-authored-by: youkaichao <[email protected]> Co-authored-by: zifeitong <[email protected]> Co-authored-by: Robert Shaw <114415538+robertgshaw2-neuralmagic@users.noreply.github.com>
Co-authored-by: Philipp Moritz <[email protected]>
andy-neuma
approved these changes
Jun 18, 2024
.github/scripts/run-tests
Outdated
| @@ -109,6 +109,10 @@ do | |||
| LOCAL_SUCCESS=0 | |||
| RESULT_XML=$(echo ${TEST} | sed -e "s/${TEST_DIR}/${RESULTS_DIR}/" | sed -e "s/.py/.xml/") | |||
|
|
|||
| # report which test is being run | |||
| # (in CI, if a test hangs, this logs *which* test is running *before* it hangs) | |||
| echo "RUNNING TEST: ${TEST}" | |||
There was a problem hiding this comment.
nice
hopefully, we can get rid of this script soon.
|
|
||
| __all__ = [ | ||
| "__version__", |
There was a problem hiding this comment.
where did the version's value go?
There was a problem hiding this comment.
Its moved to version.py + then imported
andy-neuma
approved these changes
Jun 20, 2024
.github/scripts/run-tests
Outdated
| echo "=== PASSED TEST: ${TEST} ===" | ||
| # if a file does not run any tests, pytest reports exit code of 5 | ||
| # since we skip full modules in our skipping strategy, this is common | ||
| elif [[ $LOCAL_SUCCESS == 5 ]]; then, |
There was a problem hiding this comment.
yeah ... maybe we just echo "=== SKIPPED TEST: ${TEST} ===" ?
There was a problem hiding this comment.
updated
derekk-nm
pushed a commit
that referenced
this pull request
Jun 24, 2024
Upstream sync 2024 06 16 (#310) - v0.5.0.post of vllm SUMMARY: * Merge commits from vllm-project@8f89d72 to vllm-project@0f0d8bc * Limit numpy to < 2.0 * Updated `run-tests` to print name of the test that is about to run (for debugging what hangs in automation) * Disable usage stats in automation * Temporarily disable ENTRYPOINTS (to be re-enabled in Andy's single whl PR) * Updated `run-tests` to consider exit code 5 from pytest to be a pass (since exit code 5 from pytest means that we did not run any tests) Note that vllm-project@8f89d72 is NOT included in this merge. COMPARE vs UPSTREAM: https://github.com/neuralmagic/nm-vllm/compare/upstream-sync-2024-06-16..vllm-project:vllm:v0.5.0.post1 --------- Signed-off-by: kevin <[email protected]> Signed-off-by: Travis Johnson <[email protected]> Signed-off-by: Wang, Yi A <[email protected]> Co-authored-by: Michael Goin <[email protected]> Co-authored-by: Simon Mo <[email protected]> Co-authored-by: SangBin Cho <[email protected]> Co-authored-by: Woosuk Kwon <[email protected]> Co-authored-by: Li, Jiang <[email protected]> Co-authored-by: Kevin H. Luu <[email protected]> Co-authored-by: Cody Yu <[email protected]> Co-authored-by: Arthur Kim <[email protected]> Co-authored-by: Travis Johnson <[email protected]> Co-authored-by: Sanger Steel <[email protected]> Co-authored-by: Roger Wang <[email protected]> Co-authored-by: youkaichao <[email protected]> Co-authored-by: Cyrus Leung <[email protected]> Co-authored-by: Isotr0py <[email protected]> Co-authored-by: Wang, Yi <[email protected]> Co-authored-by: Dipika Sikka <[email protected]> Co-authored-by: wenyujin333 <[email protected]> Co-authored-by: Jianan Gu <[email protected]> Co-authored-by: Tyler Michael Smith <[email protected]> Co-authored-by: zifeitong <[email protected]> Co-authored-by: Philipp Moritz <[email protected]> Co-authored-by: Antoni Baum <[email protected]> Co-authored-by: Jie Fu (傅杰) <[email protected]> Co-authored-by: Allen.Dou <[email protected]>
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Upstream sync 2024 06 16 (#310) - v0.5.0.post of vllm
SUMMARY:
run-teststo print name of the test that is about to run (for debugging what hangs in automation)run-teststo consider exit code 5 from pytest to be a pass (since exit code 5 from pytest means that we did not run any tests)Note that vllm-project@8f89d72 is NOT included in this merge.
COMPARE vs UPSTREAM:
https://github.com/neuralmagic/nm-vllm/compare/upstream-sync-2024-06-16..vllm-project:vllm:v0.5.0.post1