Parallelize CircleCI Tests #355

yuanandonly · 2022-07-26T23:52:54Z

What does this PR do?

Parallelizes CircleCI tests

Before submitting

Did you have fun?
- Make sure you had fun coding 🙃
Did you read the contributor guideline?
Was this discussed/approved via a Github issue? (no need for typos, doc improvements)
- N/A
Did you make sure to update the docs?
- N/A
Did you write any new necessary tests?
- N/A
Did you update the changelog? (if needed)
- [] N/A

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

codecov-commenter · 2022-07-27T00:54:21Z

Codecov Report

Merging #355 (d66b7ab) into main (3a7b713) will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #355      +/-   ##
==========================================
+ Coverage   93.93%   93.95%   +0.02%     
==========================================
  Files          70       70              
  Lines        3988     3988              
==========================================
+ Hits         3746     3747       +1     
+ Misses        242      241       -1

Flag	Coverage Δ
Python	`93.95% <ø> (+0.02%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
xformers/triton/layer_norm.py	`90.47% <0.00%> (+1.58%)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3a7b713...d66b7ab. Read the comment docs.

yuanandonly · 2022-07-27T16:34:52Z

Currently I have set the parallelism parameter to 4, but could be changed later on

dianaml0

Thanks for finding an opportunity to speed up the tests!

fmassa

Interesting approach!

I'm wondering if you've considered using pytest-xdist to parallelize the tests. You could then run it with pytest -n <num_jobs> tests.

Do you have thoughts on this?

* Optimize backward by 15% by using equivalent formulation * Unify everything into single kernel * Remove unused implementation * clang-format * Remove unused tensor

* Enable masking in memory-efficient attention (#333) * Add attention bias in memory-efficient attention * Add gradient for attn_mask support * Add CPU implementation * clang-format * Add benchmark scripts * Add extra loop in benchmarks * Move zeros array out of helper function * clang-format * Enable dropout in memory-efficient attention (#334) * Merge compute_scaling_coeffs and update_scaling_coeffs into a single function It wasn't needed to break it in two functions to begin with * Add CUDA implementation for dropout * clang-format * Make p be drop probability * Only CUDA supports dropout * Add benchmarks * Remove unused variables * Fix test * Cleanups and comments * Fix masking corner case when full block is masked (#339) * Add cutlass 2.9 - 858c735856a7f17bd33fe438ec76d3c9f0234e7f * Option to load from shared memory for PredicatedTileIterator * Add cutlass include dir * Ignore files in third-party for flake8/coverage * third-party -> third_party * Address comments * Revert some un-needed mods * Add attention_forward_generic.cu * Add tests * Fix duplicate calculations on baseline for mem efficient transformers * Always run all linters in CI * clang-format attention_forward_generic.cu * Benchmark: Add possibility to compare benchmarks * [isort] Ignore third_party * black autoformat * Black again + ignore third_party properly * black * Fix memory leak between the 2 benchmarks in backward * Exclude third_party/ without using pyproject.toml as it imposes isolated build which is a pain * Remove progress bar when finished * mypy * flake8 * Save results to shared folder in home location * run black * clang-format with 'run-clang-format.py' * Fix cutlass build for arch>=75 * Set tests precision for gradient more accurately * Fix precision margin * Revert changes to black * [feat] Fix importing xformers when not built (#351) authored-by: danthe3rd <danthe3rd@users.noreply.github.com> * Update black to 22.3.0 * Tweak precision for mem_eff_attention test * mem-efficient impl for f16 (#352) Co-authored-by: danthe3rd <danthe3rd> * Add support for f16 with tensorcores [sm70/sm75/sm80] (#354) * Add support for f16 with tensorcores * sm75 minimum for tensorcores * Run tests with CUDA_LAUNCH_BLOCKING=1 * Support sm70 properly * Disable tensorcore when not correctly aligned - and use 32bit accessors Co-authored-by: danthe3rd <danthe3rd> Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> * Optimize backward of memory-efficient attention by ~20% (#355) * Optimize backward by 15% by using equivalent formulation * Unify everything into single kernel * Remove unused implementation * clang-format * Remove unused tensor * Display results as we progress during benchmark (#357) Co-authored-by: danthe3rd <danthe3rd> * RFC: Ops dispatch (#356) * Ops dispatch * CI: Fix doc build * memory_efficient_attention raises when no implementation is available * type: ignore * Fix torch.device/str comparison * Make mypy happy Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> Co-authored-by: danthe3rd <danthe3rd> * [A100/f32] Use TensorCores for Q.K_t matmul with FastF32 (#358) * Use TensorCores for MM0 on Float as well * Use MultiStage MMA when available - change to FastF32 rather than FastF16 * Better alignment calculation * Just use regular f32, no fastf32 * Hackfix to handle alignment * HeuristicsMM0 -> GemmTypeQK * No longer use f16 for matmul * Add some doc * Typo * Fix build <sm80 * Alignment check based on current device compute capability * Use TORCH_INTERNAL_ASSERT Co-authored-by: danthe3rd <danthe3rd> * FlashAttention implem and dispatch (#360) * FlashAttention implem WIP * Fix flashattention forward+backward * Fix forward/backward for FlashAttention * Enable tests (more permissive) for f16 backward * Fix CI * flashattn only supports Sm75 and above * Fix CI2 * Disable K=128 when below sm80 for flashattn Co-authored-by: danthe3rd <danthe3rd> * Misc performance improvements for generic mem-efficient attention (#361) * 3% speedup by calculating mi from registers * Also compute m_prime/s_prime and exponentiate from registers * Support for Simt tiles * Fix TensorOp for V100 * Fix for A100 * Fix Simt alignment calculation * clang-format * WarpReduction before atomic call for Simt Co-authored-by: danthe3rd <danthe3rd> Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> * Update flashattention to support bf16 (#363) * Update flashattention to support bf16 * bfloat16 only on sm80 and above Co-authored-by: danthe3rd <danthe3rd> * Flashattn causal (#364) * Implement causal memory-efficient attention with FlashAttention * Update benchmarks * Fix mypy Co-authored-by: danthe3rd <danthe3rd> * Option to disable flashattention (long to build) (#362) * Option to disable flashattention (long to build) * Update setup.py Co-authored-by: danthe3rd <danthe3rd> * Remove code duplicate in attention_scaling_coefs_updater.h (#367) Co-authored-by: danthe3rd <danthe3rd> * Update .gitmodules (#366) * MemoryEff attention forward: Properly fuse matmul and enable TensorCores on the second matmul (#368) * Generic backwards * Guard backward to sm75 only * bounds checking for gradV * clang-format * Fused gemm working for Sm80/Sm75 f16/f32 * WIP * Volta TensorOp for f16 * Working on A100 again * SIMT working * Code cleanup 1 * Code cleanup2 * BUGFIX for shared memory limit * Remove code * clang-format * Remove code again * Remove draft of backward * Enforce alignment for fp16 * Fix tests * Fix constraint on seq length when not using tensorcores * Fix alignment requirements for V100/tensorcores * Clang-format * Update xformers/components/attention/csrc/cuda/attention_forward_generic.cu Co-authored-by: Francisco Massa <fvsmassa@gmail.com> * Address comments from fmassa Co-authored-by: danthe3rd <danthe3rd> Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> Co-authored-by: Francisco Massa <fvsmassa@gmail.com> * Update install instructions with submodule (#365) * Generic backward implem with cutlass (#371) * Old bw code * P100: gradV working * gk/gq working (at least for small values of M, and on P100/f16) * Further restrict supported values for bw * Fix storage into smem for Simt * More tooling for pruint/debug * Remove tests we dont need for now * Tests pass on P100 :D * 4 warps per block * Restraint on q length * Use tensorcores on V100 for f16 * Support dynamic smem for bw * Handle alignment and different dtype/arch * Fix NaNS by initializing shared memory * bw.py * Fix launch bounds * Faster 'computeDi' * minus_lse can operate on arrays * Output number of regs used etc... * Code cleanup * Hackfix for alignment check during forward * zFill to avoid nans in Sm80 + fix launch bounds * COde cleanup1 * clang-format * Fix tests * Add benchmark for K=64 Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> Co-authored-by: danthe3rd <danthe3rd> * Cutlass as submodule (#375) * Make cutlass be back at 858c735856a7f17bd33fe438ec76d3c9f0234e7f * Remove cutlass * Update submodules * Add submodule (properly) * spaces / tab * Make submodule init be recursive * Fix bad rebase * Bump tolerance for backward (#377) * Add verbose flag to CI builds (#376) * Add verbose flag to CI builds * Spurious change to rebuild cache * Add ninja * Ninja wasn't visible before, install through conda * Debugging * Source env * One more try * Forgot to uncomment a line * Another try * Cleanup * Fix for FlashAttention dispatch It requires device capability >= 7.5 * Remove generated file * Address some reviewer feedback Remove unused function and typo fix * Perf improvement on backward (#378) * Fast again on V100 * Fix correctness - missing syncthreads * Get rid of AttentionInfo Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> Co-authored-by: danthe3rd <danthe3rd@users.noreply.github.com> Co-authored-by: dan_the_3rd <43445237+danthe3rd@users.noreply.github.com>

…arch#355) * Optimize backward by 15% by using equivalent formulation * Unify everything into single kernel * Remove unused implementation * clang-format * Remove unused tensor

Chris Yuan added 2 commits July 26, 2022 16:46

test split test1

d058ab3

parallelism command fix

4128083

facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Jul 26, 2022

fix error, wip

a3c03a1

Chris Yuan added 2 commits July 26, 2022 18:10

saves artifacts

8422db8

rerun to test splitting tests by time

ef5741c

yuanandonly requested review from blefaudeux and dianaml0 July 27, 2022 16:34

dianaml0 requested a review from fmassa July 27, 2022 18:45

dianaml0 approved these changes Jul 27, 2022

View reviewed changes

yuanandonly added 5 commits July 27, 2022 14:55

Updated config.yml

ff2610d

Updated config.yml

1092aad

Updated config.yml

6b5d8f3

Updated config.yml

8bd1934

Fixing circleci config to split by test timing

1c74c17

yuanandonly force-pushed the circleci_parallelism branch from 3558e72 to 1c74c17 Compare July 28, 2022 08:18

yuanandonly added 4 commits July 28, 2022 06:06

Updated config.yml

47d47d0

Updated config.yml

d66b7ab

Updated config.yml

dd6ab21

Updated config.yml

f73b099

fmassa reviewed Jul 28, 2022

View reviewed changes

danthe3rd closed this Mar 6, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Parallelize CircleCI Tests #355

Parallelize CircleCI Tests #355

yuanandonly commented Jul 26, 2022

codecov-commenter commented Jul 27, 2022 •

edited

Loading

yuanandonly commented Jul 27, 2022

dianaml0 left a comment

fmassa left a comment

Parallelize CircleCI Tests #355

Parallelize CircleCI Tests #355

Conversation

yuanandonly commented Jul 26, 2022

What does this PR do?

Before submitting

PR review

codecov-commenter commented Jul 27, 2022 • edited Loading

Codecov Report

yuanandonly commented Jul 27, 2022

dianaml0 left a comment

Choose a reason for hiding this comment

fmassa left a comment

Choose a reason for hiding this comment

codecov-commenter commented Jul 27, 2022 •

edited

Loading