[NVFuser] Upstream push 0809 by jjsjann123 · Pull Request #83067 · pytorch/pytorch

jjsjann123 · 2022-08-09T12:54:38Z

Stack from ghstack (oldest at bottom):

-> [NVFuser] Upstream push 0809 #83067

Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/

Code changes includes:

codegen improvements:
1. removes un-necessary sync from redundant thread compute analysis
2. symmetric API for BestEffortReplay
3. support merge on trivial reductions
4. Ampere async copy improvements
bug fixes:
1. vectorization bug fixes
2. type inference patch : fixes upstream Supports infinite symbolic diff for gelu #81725
3. segmenter bug fix with deterministic iteration ordering
parser update
1. added leaky_relu
scheduler
1. normalization scheduler clean up.
2. simplifies matmul scheduling with new transform propagator
3. merge all dimensions in PW scheduler
4. various gemm related improvements
debuggability
1. nsight compute support
2. debug dump for InlinePropagator
3. Add UnaryOpType::Print

Squashed commits to WAR github API
Commits that's actually in this PR from the devel branch:

dfe02f3faed4c64477e5f5c678f21f33415d0195 Merge remote-tracking branch 'csarofeen/devel' into HEAD
16173732ecfafc4797e93c2449cfb778015a6c7a Add `TensorViewBuilder::shape(std::vector<Val*> shape)` (#1884)
7cfb7796bdcf055eb61d600b7b5c9df292950290 Merge pull request #1887 from csarofeen/upstream_merge_0803
3399f6de62061d30781de50ef1862bbfb1615173 Merge remote-tracking branch 'origin/viable/strict' into HEAD
01208f5bba3bc158d41ccbefa0ee2c5ceea7aedb Add `UnaryOpType::Print` which can be helpful for debugging (#1878)
0646522454aa715ef164c88a73fb8bdddc706805 Remove redundant TORCH_INTERNAL_ASSERT in lower_magic_zero.cpp (#1881)
7bc76aa219293a59e4166e258d76289fe13633ca Fix most inlined propagator for mismatched dims (#1875)
501f4aa270bf4dd47b0d2f4860bc6f23ebc32a38 Nonaffine swizzle formulation ep.2: Loop swizzle variant. (#1826)
d863d690f923047a85b5229a787118708f810741 Ampere async copy ep.2: circular buffering extension to support pipelined matmul operand load (#1827)
e0ae11a61c87cd998e88ddd79a496548171c31e0 Larger sized mma instructions to support full vectorization (#1824)
9bb4cf7a66b098f04c9d95a2d34ab2bceee151b3 fragment iteration to support fully unrolled mma ops (#1823)
a48270a18dc2d3accc2626758d14d5858ae55032 Merge all dims in pointwise scheduler (#1872)
172fb3673fb4aaf4c1e889922a4fc5c06cbd59f7 Make MostInlined and BestEffort inline propagation no longer assert replayed (#1868)
a64462a5ac2fcf57a177bf36b0f26c61a4e252a4 Allow trivial reduction to be merged (#1871)
440102bcda6eb1dcd42d5fa5aeab9d6b049956bc Symmetric API for BestEffortReplay (#1870)
d1caf330c08ea8002f7133ca655bbd5b28c4eb98 Some misc cleanups/refactor split out from #1854 (#1867)
1013eda50be38eac96c00ba781340ac199d5a136 Remove some welford specific logic. (#1864)
51589d36be5a101d06e641fe0400b39028b7cb81 Some cleanups on tests and heuristics params (#1866)
a6b3e70da5dee51dbc246347228ea21384e46ac3 Segmenter bug fix, and deterministic iteration ordering.  (#1865)
1b665b9b5e562d6f0caba5e7319e83e5df64104f Add nullptr checks to IrBuilder (#1861)
1cd9451d7493f631c2837ba07c1ea93a74e83a15 Simplify matmul scheduling with the new transform propagator.  (#1817)
bbc1fb9b8c454f557ab9fcf5b1c3cef9b9e136d0 Add leaky_relu operation (#1852)
e842a9bab5e9f7289b7ce33ee37a682b22373f49 Minor cleanup in pointwise scheduler (#1858)
9ee850ca2f7f51dd5269bffb1255e485f809282d Fix stringstream usage (#1857)
20a36c1e4f28c4ff9837e56784be2686d17435f3 Improve nsight compute support (#1855)
405910308301097297b55c34d560aab6a360e897 Remove debugging `true ||` from getPointwiseHeuristics (#1822)
01117bfe8fdfacdbfdcfba9a624cdf900fe044d4 Misc cleanup (#1853)
5cc64943dc381a568223140bce0f22163c01e29f Apply the magic-zero protection to each indexed domain individually for predicate indexing (#1846)
92e6f0207e3a89fe90fd5cd3ffc575dfd766ba00 Cleanup normalization scheduler (#1845)
db89c6591a2f21130599a93675e0615e55564e41 Type inference patch (#1848)
102fe93a4605ca465cda26ebaee4ba1af2026901 Add debug dump for InlinePropagator (#1847)
b7a4d93d375a6e2ddef483763c93ffddc62ec452 Redundant thread compute analysis to avoid un-necessary sync insertion (#1687)
942be5b256056d0e02877361b814ae6af32ca15f Upstream ci build fixes (#1842)
0b83645915029d67f9345aa4649b8c6f62b0061b Fix vectorization bug introduced in #1831 (#1840)
63630f1ae091180e541932a9d9dc598e0a9902dd Move MaxProducerPosUpdater into InlinePropagator::tearDown (#1825)
9135a963c01d97ba34b1a7d2f106e78a13fd6651 Fix transpose benchmark dtype (#1839)
2c9a6c02312d5bf4f83cde653b847b4f85849432 Add extra configurability to `parallelizeAllLike` (#1831)

RUN_TORCHBENCH: nvfuser

Differential Revision: D38543000

Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Code changes includes: - codegen improvements: 1. removes un-necessary sync from redundant thread compute analysis 2. symmetric API for BestEffortReplay 3. support merge on trivial reductions 4. Ampere async copy improvements - bug fixes: 1. vectorization bug fixes 2. type inference patch : fixes upstream #81725 3. segmenter bug fix with deterministic iteration ordering - parser update 1. added leaky_relu - scheduler 1. normalization scheduler clean up. 2. simplifies matmul scheduling with new transform propagator 3. merge all dimensions in PW scheduler 4. various gemm related improvements - debuggability 1. nsight compute support 2. debug dump for InlinePropagator 3. Add `UnaryOpType::Print` Squashed commits to WAR github API Commits that's actually in this PR from the devel branch: ``` dfe02f3 Merge remote-tracking branch 'csarofeen/devel' into HEAD 1617373 Add `TensorViewBuilder::shape(std::vector<Val*> shape)` (#1884) 7cfb779 Merge pull request #1887 from csarofeen/upstream_merge_0803 3399f6d Merge remote-tracking branch 'origin/viable/strict' into HEAD 01208f5 Add `UnaryOpType::Print` which can be helpful for debugging (#1878) 0646522 Remove redundant TORCH_INTERNAL_ASSERT in lower_magic_zero.cpp (#1881) 7bc76aa Fix most inlined propagator for mismatched dims (#1875) 501f4aa Nonaffine swizzle formulation ep.2: Loop swizzle variant. (#1826) d863d69 Ampere async copy ep.2: circular buffering extension to support pipelined matmul operand load (#1827) e0ae11a Larger sized mma instructions to support full vectorization (#1824) 9bb4cf7 fragment iteration to support fully unrolled mma ops (#1823) a48270a Merge all dims in pointwise scheduler (#1872) 172fb36 Make MostInlined and BestEffort inline propagation no longer assert replayed (#1868) a64462a Allow trivial reduction to be merged (#1871) 440102b Symmetric API for BestEffortReplay (#1870) d1caf33 Some misc cleanups/refactor split out from #1854 (#1867) 1013eda Remove some welford specific logic. (#1864) 51589d3 Some cleanups on tests and heuristics params (#1866) a6b3e70 Segmenter bug fix, and deterministic iteration ordering. (#1865) 1b665b9 Add nullptr checks to IrBuilder (#1861) 1cd9451 Simplify matmul scheduling with the new transform propagator. (#1817) bbc1fb9 Add leaky_relu operation (#1852) e842a9b Minor cleanup in pointwise scheduler (#1858) 9ee850c Fix stringstream usage (#1857) 20a36c1 Improve nsight compute support (#1855) 4059103 Remove debugging `true ||` from getPointwiseHeuristics (#1822) 01117bf Misc cleanup (#1853) 5cc6494 Apply the magic-zero protection to each indexed domain individually for predicate indexing (#1846) 92e6f02 Cleanup normalization scheduler (#1845) db89c65 Type inference patch (#1848) 102fe93 Add debug dump for InlinePropagator (#1847) b7a4d93 Redundant thread compute analysis to avoid un-necessary sync insertion (#1687) 942be5b Upstream ci build fixes (#1842) 0b83645 Fix vectorization bug introduced in #1831 (#1840) 63630f1 Move MaxProducerPosUpdater into InlinePropagator::tearDown (#1825) 9135a96 Fix transpose benchmark dtype (#1839) 2c9a6c0 Add extra configurability to `parallelizeAllLike` (#1831) ``` RUN_TORCHBENCH: nvfuser [ghstack-poisoned]

facebook-github-bot · 2022-08-09T12:54:46Z

🔗 Helpful links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/83067
📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓Need help or want to give feedback on the CI? Visit our office hours

✅ No Failures (0 Pending)

As of commit 34c2149 (more details on the Dr. CI page):

Expand to see more

💚 💚 Looks good so far! There are no failures yet. 💚 💚

This comment was automatically generated by Dr. CI (expand for details).

Please report bugs/suggestions to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Code changes includes: - codegen improvements: 1. removes un-necessary sync from redundant thread compute analysis 2. symmetric API for BestEffortReplay 3. support merge on trivial reductions 4. Ampere async copy improvements - bug fixes: 1. vectorization bug fixes 2. type inference patch : fixes upstream #81725 3. segmenter bug fix with deterministic iteration ordering - parser update 1. added leaky_relu - scheduler 1. normalization scheduler clean up. 2. simplifies matmul scheduling with new transform propagator 3. merge all dimensions in PW scheduler 4. various gemm related improvements - debuggability 1. nsight compute support 2. debug dump for InlinePropagator 3. Add `UnaryOpType::Print` Squashed commits to WAR github API Commits that's actually in this PR from the devel branch: ``` dfe02f3 Merge remote-tracking branch 'csarofeen/devel' into HEAD 1617373 Add `TensorViewBuilder::shape(std::vector<Val*> shape)` (#1884) 7cfb779 Merge pull request #1887 from csarofeen/upstream_merge_0803 3399f6d Merge remote-tracking branch 'origin/viable/strict' into HEAD 01208f5 Add `UnaryOpType::Print` which can be helpful for debugging (#1878) 0646522 Remove redundant TORCH_INTERNAL_ASSERT in lower_magic_zero.cpp (#1881) 7bc76aa Fix most inlined propagator for mismatched dims (#1875) 501f4aa Nonaffine swizzle formulation ep.2: Loop swizzle variant. (#1826) d863d69 Ampere async copy ep.2: circular buffering extension to support pipelined matmul operand load (#1827) e0ae11a Larger sized mma instructions to support full vectorization (#1824) 9bb4cf7 fragment iteration to support fully unrolled mma ops (#1823) a48270a Merge all dims in pointwise scheduler (#1872) 172fb36 Make MostInlined and BestEffort inline propagation no longer assert replayed (#1868) a64462a Allow trivial reduction to be merged (#1871) 440102b Symmetric API for BestEffortReplay (#1870) d1caf33 Some misc cleanups/refactor split out from #1854 (#1867) 1013eda Remove some welford specific logic. (#1864) 51589d3 Some cleanups on tests and heuristics params (#1866) a6b3e70 Segmenter bug fix, and deterministic iteration ordering. (#1865) 1b665b9 Add nullptr checks to IrBuilder (#1861) 1cd9451 Simplify matmul scheduling with the new transform propagator. (#1817) bbc1fb9 Add leaky_relu operation (#1852) e842a9b Minor cleanup in pointwise scheduler (#1858) 9ee850c Fix stringstream usage (#1857) 20a36c1 Improve nsight compute support (#1855) 4059103 Remove debugging `true ||` from getPointwiseHeuristics (#1822) 01117bf Misc cleanup (#1853) 5cc6494 Apply the magic-zero protection to each indexed domain individually for predicate indexing (#1846) 92e6f02 Cleanup normalization scheduler (#1845) db89c65 Type inference patch (#1848) 102fe93 Add debug dump for InlinePropagator (#1847) b7a4d93 Redundant thread compute analysis to avoid un-necessary sync insertion (#1687) 942be5b Upstream ci build fixes (#1842) 0b83645 Fix vectorization bug introduced in #1831 (#1840) 63630f1 Move MaxProducerPosUpdater into InlinePropagator::tearDown (#1825) 9135a96 Fix transpose benchmark dtype (#1839) 2c9a6c0 Add extra configurability to `parallelizeAllLike` (#1831) ``` RUN_TORCHBENCH: nvfuser ghstack-source-id: 3745722 Pull Request resolved: #83067

jjsjann123 · 2022-08-09T12:56:47Z

rebased from #82937. I saw a few time out on macos there, fixes all other failures. 🤞 Hopefully this one goes smoother.

cc'ing @davidberard98

davidberard98 · 2022-08-09T17:20:47Z

@davidberard98 has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

jjsjann123 · 2022-08-10T08:01:46Z

Is there any issues coming from internal builds?

I'm asking because we have a couple PR that we want to push to upstream for some dropout related failures. I'm debating on whether rebasing them into this one, or maybe cherry-pick those after we have this PR merged.

davidberard98 · 2022-08-10T14:55:50Z

@jjsjann123 it's looking okay for internal builds, I will try to land today - let's wait to see if it passes before we merge anything else into this PR

jjsjann123 · 2022-08-10T16:05:52Z

@jjsjann123 it's looking okay for internal builds, I will try to land today - let's wait to see if it passes before we merge anything else into this PR

Sounds good to me~ If this one is good to land let's not touch it any more. I'll start a separate PR right after it's merged.

facebook-github-bot · 2022-08-10T21:01:35Z

@pytorchbot merge

(Initiating merge automatically since Phabricator Diff has merged)

pytorchmergebot · 2022-08-10T21:02:51Z

@pytorchbot successfully started a merge job. Check the current status here

github-actions · 2022-08-10T21:03:35Z

Hey @jjsjann123.
You've committed this PR, but it does not have both a 'release notes: ...' and 'topics: ...' label. Please add one of each to the PR. The 'release notes: ...' label should represent the part of PyTorch that this PR changes (fx, autograd, distributed, etc) and the 'topics: ...' label should represent the kind of PR it is (not user facing, new feature, bug fix, perf improvement, etc). The list of valid labels can be found here for the 'release notes: ...' and here for the 'topics: ...'.
For changes that are 'topic: not user facing' there is no need for a release notes label.

Summary: Pull Request resolved: #83067 Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Code changes includes: - codegen improvements: 1. removes un-necessary sync from redundant thread compute analysis 2. symmetric API for BestEffortReplay 3. support merge on trivial reductions 4. Ampere async copy improvements - bug fixes: 1. vectorization bug fixes 2. type inference patch : fixes upstream #81725 3. segmenter bug fix with deterministic iteration ordering - parser update 1. added leaky_relu - scheduler 1. normalization scheduler clean up. 2. simplifies matmul scheduling with new transform propagator 3. merge all dimensions in PW scheduler 4. various gemm related improvements - debuggability 1. nsight compute support 2. debug dump for InlinePropagator 3. Add `UnaryOpType::Print` Squashed commits to WAR github API Commits that's actually in this PR from the devel branch: ``` dfe02f3 Merge remote-tracking branch 'csarofeen/devel' into HEAD 1617373 Add `TensorViewBuilder::shape(std::vector<Val*> shape)` (#1884) 7cfb779 Merge pull request #1887 from csarofeen/upstream_merge_0803 3399f6d Merge remote-tracking branch 'origin/viable/strict' into HEAD 01208f5 Add `UnaryOpType::Print` which can be helpful for debugging (#1878) 0646522 Remove redundant TORCH_INTERNAL_ASSERT in lower_magic_zero.cpp (#1881) 7bc76aa Fix most inlined propagator for mismatched dims (#1875) 501f4aa Nonaffine swizzle formulation ep.2: Loop swizzle variant. (#1826) d863d69 Ampere async copy ep.2: circular buffering extension to support pipelined matmul operand load (#1827) e0ae11a Larger sized mma instructions to support full vectorization (#1824) 9bb4cf7 fragment iteration to support fully unrolled mma ops (#1823) a48270a Merge all dims in pointwise scheduler (#1872) 172fb36 Make MostInlined and BestEffort inline propagation no longer assert replayed (#1868) a64462a Allow trivial reduction to be merged (#1871) 440102b Symmetric API for BestEffortReplay (#1870) d1caf33 Some misc cleanups/refactor split out from #1854 (#1867) 1013eda Remove some welford specific logic. (#1864) 51589d3 Some cleanups on tests and heuristics params (#1866) a6b3e70 Segmenter bug fix, and deterministic iteration ordering. (#1865) 1b665b9 Add nullptr checks to IrBuilder (#1861) 1cd9451 Simplify matmul scheduling with the new transform propagator. (#1817) bbc1fb9 Add leaky_relu operation (#1852) e842a9b Minor cleanup in pointwise scheduler (#1858) 9ee850c Fix stringstream usage (#1857) 20a36c1 Improve nsight compute support (#1855) 4059103 Remove debugging `true ||` from getPointwiseHeuristics (#1822) 01117bf Misc cleanup (#1853) 5cc6494 Apply the magic-zero protection to each indexed domain individually for predicate indexing (#1846) 92e6f02 Cleanup normalization scheduler (#1845) db89c65 Type inference patch (#1848) 102fe93 Add debug dump for InlinePropagator (#1847) b7a4d93 Redundant thread compute analysis to avoid un-necessary sync insertion (#1687) 942be5b Upstream ci build fixes (#1842) 0b83645 Fix vectorization bug introduced in #1831 (#1840) 63630f1 Move MaxProducerPosUpdater into InlinePropagator::tearDown (#1825) 9135a96 Fix transpose benchmark dtype (#1839) 2c9a6c0 Add extra configurability to `parallelizeAllLike` (#1831) ``` RUN_TORCHBENCH: nvfuser Test Plan: Imported from OSS Reviewed By: malfet Differential Revision: D38543000 Pulled By: davidberard98 fbshipit-source-id: 752edbfbced14fe01b84e417f23cc941b2148842

davidberard98 · 2022-08-10T22:46:16Z

@jjsjann123 btw, seeing some small regressions in torchbench ci, 7% on autogen-34, 5% on autogen-41. Could very well be typical variations in runtime, but perhaps you can take a look if you get a chance? (also seeing ~6% improvement on autogen-40, autogen-43)

jjsjann123 · 2022-08-11T00:13:38Z

@jjsjann123 btw, seeing some small regressions in torchbench ci, 7% on autogen-34, 5% on autogen-41. Could very well be typical variations in runtime, but perhaps you can take a look if you get a chance? (also seeing ~6% improvement on autogen-40, autogen-43)

Thanks for double checking the regression. I'll take a look when I got some bandwidth~

cherry-pick upstream CI fixes from pytorch#83067 & pytorch#83239

Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Code changes includes: - codegen improvements: 1. removes un-necessary sync from redundant thread compute analysis 2. symmetric API for BestEffortReplay 3. support merge on trivial reductions 4. Ampere async copy improvements - bug fixes: 1. vectorization bug fixes 2. type inference patch : fixes upstream #81725 3. segmenter bug fix with deterministic iteration ordering - parser update 1. added leaky_relu - scheduler 1. normalization scheduler clean up. 2. simplifies matmul scheduling with new transform propagator 3. merge all dimensions in PW scheduler 4. various gemm related improvements - debuggability 1. nsight compute support 2. debug dump for InlinePropagator 3. Add `UnaryOpType::Print` Squashed commits to WAR github API Commits that's actually in this PR from the devel branch: ``` dfe02f3faed4c64477e5f5c678f21f33415d0195 Merge remote-tracking branch 'csarofeen/devel' into HEAD 16173732ecfafc4797e93c2449cfb778015a6c7a Add `TensorViewBuilder::shape(std::vector<Val*> shape)` (#1884) 7cfb7796bdcf055eb61d600b7b5c9df292950290 Merge pull request #1887 from csarofeen/upstream_merge_0803 3399f6de62061d30781de50ef1862bbfb1615173 Merge remote-tracking branch 'origin/viable/strict' into HEAD 01208f5bba3bc158d41ccbefa0ee2c5ceea7aedb Add `UnaryOpType::Print` which can be helpful for debugging (#1878) 0646522454aa715ef164c88a73fb8bdddc706805 Remove redundant TORCH_INTERNAL_ASSERT in lower_magic_zero.cpp (#1881) 7bc76aa219293a59e4166e258d76289fe13633ca Fix most inlined propagator for mismatched dims (#1875) 501f4aa270bf4dd47b0d2f4860bc6f23ebc32a38 Nonaffine swizzle formulation ep.2: Loop swizzle variant. (#1826) d863d690f923047a85b5229a787118708f810741 Ampere async copy ep.2: circular buffering extension to support pipelined matmul operand load (#1827) e0ae11a61c87cd998e88ddd79a496548171c31e0 Larger sized mma instructions to support full vectorization (#1824) 9bb4cf7a66b098f04c9d95a2d34ab2bceee151b3 fragment iteration to support fully unrolled mma ops (#1823) a48270a18dc2d3accc2626758d14d5858ae55032 Merge all dims in pointwise scheduler (#1872) 172fb3673fb4aaf4c1e889922a4fc5c06cbd59f7 Make MostInlined and BestEffort inline propagation no longer assert replayed (#1868) a64462a5ac2fcf57a177bf36b0f26c61a4e252a4 Allow trivial reduction to be merged (#1871) 440102bcda6eb1dcd42d5fa5aeab9d6b049956bc Symmetric API for BestEffortReplay (#1870) d1caf330c08ea8002f7133ca655bbd5b28c4eb98 Some misc cleanups/refactor split out from #1854 (#1867) 1013eda50be38eac96c00ba781340ac199d5a136 Remove some welford specific logic. (#1864) 51589d36be5a101d06e641fe0400b39028b7cb81 Some cleanups on tests and heuristics params (#1866) a6b3e70da5dee51dbc246347228ea21384e46ac3 Segmenter bug fix, and deterministic iteration ordering. (#1865) 1b665b9b5e562d6f0caba5e7319e83e5df64104f Add nullptr checks to IrBuilder (#1861) 1cd9451d7493f631c2837ba07c1ea93a74e83a15 Simplify matmul scheduling with the new transform propagator. (#1817) bbc1fb9b8c454f557ab9fcf5b1c3cef9b9e136d0 Add leaky_relu operation (#1852) e842a9bab5e9f7289b7ce33ee37a682b22373f49 Minor cleanup in pointwise scheduler (#1858) 9ee850ca2f7f51dd5269bffb1255e485f809282d Fix stringstream usage (#1857) 20a36c1e4f28c4ff9837e56784be2686d17435f3 Improve nsight compute support (#1855) 405910308301097297b55c34d560aab6a360e897 Remove debugging `true ||` from getPointwiseHeuristics (#1822) 01117bfe8fdfacdbfdcfba9a624cdf900fe044d4 Misc cleanup (#1853) 5cc64943dc381a568223140bce0f22163c01e29f Apply the magic-zero protection to each indexed domain individually for predicate indexing (#1846) 92e6f0207e3a89fe90fd5cd3ffc575dfd766ba00 Cleanup normalization scheduler (#1845) db89c6591a2f21130599a93675e0615e55564e41 Type inference patch (#1848) 102fe93a4605ca465cda26ebaee4ba1af2026901 Add debug dump for InlinePropagator (#1847) b7a4d93d375a6e2ddef483763c93ffddc62ec452 Redundant thread compute analysis to avoid un-necessary sync insertion (#1687) 942be5b256056d0e02877361b814ae6af32ca15f Upstream ci build fixes (#1842) 0b83645915029d67f9345aa4649b8c6f62b0061b Fix vectorization bug introduced in #1831 (#1840) 63630f1ae091180e541932a9d9dc598e0a9902dd Move MaxProducerPosUpdater into InlinePropagator::tearDown (#1825) 9135a963c01d97ba34b1a7d2f106e78a13fd6651 Fix transpose benchmark dtype (#1839) 2c9a6c02312d5bf4f83cde653b847b4f85849432 Add extra configurability to `parallelizeAllLike` (#1831) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D38543000](https://our.internmc.facebook.com/intern/diff/D38543000) Pull Request resolved: pytorch/pytorch#83067 Approved by: https://github.com/davidberard98

Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Code changes includes: - codegen improvements: 1. removes un-necessary sync from redundant thread compute analysis 2. symmetric API for BestEffortReplay 3. support merge on trivial reductions 4. Ampere async copy improvements - bug fixes: 1. vectorization bug fixes 2. type inference patch : fixes upstream #81725 3. segmenter bug fix with deterministic iteration ordering - parser update 1. added leaky_relu - scheduler 1. normalization scheduler clean up. 2. simplifies matmul scheduling with new transform propagator 3. merge all dimensions in PW scheduler 4. various gemm related improvements - debuggability 1. nsight compute support 2. debug dump for InlinePropagator 3. Add `UnaryOpType::Print` Squashed commits to WAR github API Commits that's actually in this PR from the devel branch: ``` dfe02f3faed4c64477e5f5c678f21f33415d0195 Merge remote-tracking branch 'csarofeen/devel' into HEAD 1cc4977 Add `TensorViewBuilder::shape(std::vector<Val*> shape)` (#1884) b058179 Merge pull request #1887 from csarofeen/upstream_merge_0803 d61ae45 Merge remote-tracking branch 'origin/viable/strict' into HEAD 9186fa2 Add `UnaryOpType::Print` which can be helpful for debugging (#1878) 6e46c0b Remove redundant TORCH_INTERNAL_ASSERT in lower_magic_zero.cpp (#1881) c000186 Fix most inlined propagator for mismatched dims (#1875) f1071b6 Nonaffine swizzle formulation ep.2: Loop swizzle variant. (#1826) 4c6d5b8 Ampere async copy ep.2: circular buffering extension to support pipelined matmul operand load (#1827) a286262 Larger sized mma instructions to support full vectorization (#1824) 9db41cf fragment iteration to support fully unrolled mma ops (#1823) c9c8b57 Merge all dims in pointwise scheduler (#1872) b8544dc Make MostInlined and BestEffort inline propagation no longer assert replayed (#1868) 4cadfd2 Allow trivial reduction to be merged (#1871) 86de4f0 Symmetric API for BestEffortReplay (#1870) fcd1dbe Some misc cleanups/refactor split out from #1854 (#1867) 28b3d56 Remove some welford specific logic. (#1864) 4993b94 Some cleanups on tests and heuristics params (#1866) e4cee81 Segmenter bug fix, and deterministic iteration ordering. (#1865) e4158b4 Add nullptr checks to IrBuilder (#1861) fbaabe1 Simplify matmul scheduling with the new transform propagator. (#1817) b6fb58c Add leaky_relu operation (#1852) 19a2f89 Minor cleanup in pointwise scheduler (#1858) 52c858f Fix stringstream usage (#1857) 484441f Improve nsight compute support (#1855) 5067a35 Remove debugging `true ||` from getPointwiseHeuristics (#1822) f884867 Misc cleanup (#1853) 04a95b7 Apply the magic-zero protection to each indexed domain individually for predicate indexing (#1846) 30941ed Cleanup normalization scheduler (#1845) 93edf4c Type inference patch (#1848) 05bd85e Add debug dump for InlinePropagator (#1847) af085e5 Redundant thread compute analysis to avoid un-necessary sync insertion (#1687) ead1284 Upstream ci build fixes (#1842) 975ad71 Fix vectorization bug introduced in #1831 (#1840) 85494a8 Move MaxProducerPosUpdater into InlinePropagator::tearDown (#1825) 9135a963c01d97ba34b1a7d2f106e78a13fd6651 Fix transpose benchmark dtype (#1839) 8912de7 Add extra configurability to `parallelizeAllLike` (#1831) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D38543000](https://our.internmc.facebook.com/intern/diff/D38543000) Pull Request resolved: pytorch/pytorch#83067 Approved by: https://github.com/davidberard98

facebook-github-bot added the cla signed label Aug 9, 2022

facebook-github-bot added the oncall: jit Add this issue/PR to JIT oncall triage queue label Aug 9, 2022

jjsjann123 added the ciflow/trunk Trigger trunk jobs on your pull request label Aug 9, 2022

jjsjann123 requested review from davidberard98 and kevinstephano August 9, 2022 12:57

pytorchbot added the open source label Aug 9, 2022

This was referenced Aug 9, 2022

Type inference patch (#1848) #81792

Closed

expanded_extent_ == nullptr || isBroadcast() INTERNAL ASSERT FAILED #82858

Closed

davidberard98 approved these changes Aug 10, 2022

View reviewed changes

pytorchmergebot added the Merged label Aug 10, 2022

pytorchmergebot closed this in df741c5 Aug 10, 2022

jjsjann123 deleted the gh/jjsjann123/2/head branch August 10, 2022 21:44

jjsjann123 added a commit to csarofeen/pytorch that referenced this pull request Aug 26, 2022

cherry-pick upstream fixes from pytorch#83067

f5493d0

jjsjann123 added a commit to csarofeen/pytorch that referenced this pull request Aug 26, 2022

Upstream cherry pick fixes 0811 (#1934)

88de85e

cherry-pick upstream CI fixes from pytorch#83067 & pytorch#83239

atalman mentioned this pull request Feb 19, 2024

[RFC] Intel GPU ATen Operations Upstreaming Options #119682

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[NVFuser] Upstream push 0809#83067

[NVFuser] Upstream push 0809#83067
jjsjann123 wants to merge 1 commit intogh/jjsjann123/2/basefrom
gh/jjsjann123/2/head

jjsjann123 commented Aug 9, 2022 •

edited by davidberard98

Loading

Uh oh!

facebook-github-bot commented Aug 9, 2022 •

edited

Loading

Uh oh!

jjsjann123 commented Aug 9, 2022

Uh oh!

davidberard98 commented Aug 9, 2022

Uh oh!

jjsjann123 commented Aug 10, 2022

Uh oh!

davidberard98 commented Aug 10, 2022

Uh oh!

jjsjann123 commented Aug 10, 2022

Uh oh!

facebook-github-bot commented Aug 10, 2022

Uh oh!

pytorchmergebot commented Aug 10, 2022

Uh oh!

github-actions bot commented Aug 10, 2022

Uh oh!

davidberard98 commented Aug 10, 2022

Uh oh!

jjsjann123 commented Aug 11, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Conversation

jjsjann123 commented Aug 9, 2022 • edited by davidberard98 Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

facebook-github-bot commented Aug 9, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful links

✅ No Failures (0 Pending)

Uh oh!

jjsjann123 commented Aug 9, 2022

Uh oh!

davidberard98 commented Aug 9, 2022

Uh oh!

jjsjann123 commented Aug 10, 2022

Uh oh!

davidberard98 commented Aug 10, 2022

Uh oh!

jjsjann123 commented Aug 10, 2022

Uh oh!

facebook-github-bot commented Aug 10, 2022

Uh oh!

pytorchmergebot commented Aug 10, 2022

Uh oh!

github-actions bot commented Aug 10, 2022

Uh oh!

davidberard98 commented Aug 10, 2022

Uh oh!

jjsjann123 commented Aug 11, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

jjsjann123 commented Aug 9, 2022 •

edited by davidberard98

Loading

facebook-github-bot commented Aug 9, 2022 •

edited

Loading