Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Segmenter] test the groups the same order as they are merged #1949

Merged
merged 1 commit into from
Sep 1, 2022
Merged

[Segmenter] test the groups the same order as they are merged #1949

merged 1 commit into from
Sep 1, 2022

Conversation

shmsong
Copy link

@shmsong shmsong commented Sep 1, 2022

This PR is a work around for #1947. It makes sure that the input and output ordering of the merged groups stay consistent in different phases of segment pass so that canSchedule returns consistent results.

@shmsong shmsong mentioned this pull request Sep 1, 2022
Closed
@IvanYashchuk
Copy link
Collaborator

I can confirm that with these changes both code snippets from the issue description and #1947 (comment) now work.

jjsjann123
jjsjann123 reviewed Sep 1, 2022
View reviewed changes
torch/csrc/jit/codegen/cuda/fusion_segmenter.cpp
@@ -2827,7 +2857,7 @@ void SegmentCandidateFinder::findSegments() {

auto candidate_it = candidates.begin();
while (candidate_it != candidates.end() &&
!codeGenSupportedMerge(candidate_it->edge)) {
!codeGenSupportedMerge(group, candidate_it->group)) {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This change is counter-intuitive.

The old API only tries to merge all candidate connected by candidate_it->edge. Shouldn't we plumb the connection check to getMergeCandidates instead of assert on the connectivity inside codeGenSupportedMerge?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

getMergeCandidates already guarantees that they are connected by construction, since looking at immediate neighbors is exactly what it does.

the connectivity on codeGenSupportedMerge I added isn't strictly necessary but just a sanity check that edge -> group1, group2 is semantically equivalent to what we've been doing.

I'd expect removing that assert in future PR's when we try to test more candidate patterns, which hasn't been prioritized.

@csarofeen
Copy link
Owner

Which phases were inconsistent? Is this if we identify a segment is good, then we make the segment, the output ordering isn't the same because supportedMerge and makeFusion can have different ordering of outputs?

csarofeen
csarofeen approved these changes Sep 1, 2022
View reviewed changes
Copy link
Owner

@csarofeen csarofeen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this PR seems fine, if I understand correctly you're just making sure processing an edge we don't merge groups as (consumer, producer) but always merge as (producer, consumer).

@csarofeen
Copy link
Owner

This seems fine as a WAR, though we definitely should figure out what's going on in #1947.

@shmsong
Copy link
Author

shmsong commented Sep 1, 2022

Merging as is for now to unblock.

@shmsong shmsong merged commit 992e17c into devel Sep 1, 2022
@shmsong shmsong deleted the segmenter_io_order branch September 1, 2022 17:23
@csarofeen csarofeen mentioned this pull request Sep 2, 2022
Merged
3 tasks
jjsjann123 added a commit that referenced this pull request Nov 9, 2022
@jjsjann123 @pytorchmergebot
[NVFuser] Upstream push 1026 (pytorch#87779)
7b419e8 
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/

Codegen changes include:

* codegen improvement:
    i. allow non-root trivial reductions, allow empty/no-op fusion
    ii. fixes vectorization checks and size calculation
    iii. bank conflict handle improvement
    iv. enables transpose scheduler

* misc:
    i. CI tests failure fixes
    ii. cpp tests file clean up
    iii. trivial forwarding supports added in codegen runtime
    iv. added factory methods support in codegen

Commits that's in this PR from the devel branch:

```
7117a7e patching nvfuser conv cudnn test numerics mismatch (#2048)
65af1a4 Inserting sync for redundant parallel types is already done at the (#2023)
6ac74d1 Fix sync map (#2047)
f5bca33 Bank conflict checker improvements (#2032)
d2ca7e3 Minor update on cp.async code generation. (#1901)
d36cf61 Test file cleanup (#2040)
0b8e83f Allow non-root trivial reductions (#2037)
a2dfe40 Fix vectorize size calculation (#2035)
e040676 Use withPredicate to replace setPredicate to maintain Exprs immutable (#2025)
197221b removing ci workflow (#2034)
40e2703 Reduction rand like patch (#2031)
bc77266 Add utility for checking bank conflict of shared memory (#2029)
ddd1cf7 Add back FusionReductionWithTrivialReduction_CUDA (#2030)
fbd97e5 Revert "Cleanup trivial reduction workarounds (#2006)" (#2024)
bca20c1 Cleanup trivial reduction workarounds (#2006)
e4b6585 Trivial forwarding (#1995)
1a0e355 Fix contiguity analysis of predicates to match updated contiguity. (#1991)
a4effa6 Enable output allocation cache (#2010)
35440b7 Patching bn inference (#2016)
0f9f0b4 Add matmul benchmark (#2007)
45045cd Enable tests previously disabled due to an aliasing bug (#2005)
967aa77 Contiguous indexing for View operations (#1990)
a43cb20 Make inlining even more modular (#2004)
dc45835 Test util cleanup (#2003)
3ca21eb More strict validation (#2000)
a7a7d57 Fix build problem (#1999)
fc235b0 Just fixes comments (#1998)
482386c cleanup (#1997)
4cbe0db Improve divisible split detection (#1970)
42ccc52 Minor build fix. (#1996)
fcf8c09 Cleanup of lower_utils.cpp: Isolate out GpuLower usage (#1989)
15f2f6d Move ConcretizedBroadcastDomains to shared_ptr in GpuLower. (#1988)
8f1c7f5 Minor cleanup lower_unroll.cpp (#1994)
1d9858c Minor cleanup (#1992)
f262d9c Add support for uniform RNG (#1986)
eb1dad1 Remove non-const functions, remove GpuLower instance on build, pass in ca_map. (#1987)
634820c Add support for some empty fusion (#1981)
eabe8d8 Segment self mapping fusions (#1954)
e96aacf Enable Transpose operation (#1882)
425dce2 Add a null scheduler that helps segmenting away no-op schedules (#1835)
306d4a6 Fix canScheduleCompileTime check of transpose scheduler (#1969)
b1bd32c Minor fix (#1967)
bd93578 Enable transpose scheduler (#1927)
b7a206e Move scheduler vectorize utilities into their own file (#1959)
d9420e4 View scheduling (#1928)
c668e13 Upstream push ci fixes (#1965)
c40202b Fix dump effective bandwidth (#1962)
93505bc WAR on index mapping when exact and permissive maps differ (#1960)
45e95fd Allow splitting inner-most ID to create virtual innermost ID in transpose scheduler (#1930)
a3ecb33 Improve the comments at the beginning of index_compute.h (#1946)
f7bc341 Remove unused variables (#1955)
df3393a Some cleanup (#1957)
7d1d7c8 TVDomainGuard factory (#1953)
357ba22 Fill allocation with nan on tests (#1956)
8eafc54 Fix detection of unmappable root domains (#1952)
90a51f2 Some indexing cleanups, Add eye support (#1940)
ddc01e4 Exclude unsupported data types (#1951)
992e17c test the groups the same order as they are merged (#1949)
208262b Move detection of self mapping IDs to IterDomainGraph from (#1941)
ac4de38 Merge pull request #1945 from csarofeen/master_merge_0828
6310948 Add full, full_like, zeros, zeros_like, ones, ones_like (#1943)
aab10bc Merge remote-tracking branch 'upstream/viable/strict' into HEAD
4c254c0 Fix arange when step is negative (#1942)
89330aa Tensor factories must set the output shape as its input (#1939)
```

RUN_TORCHBENCH: nvfuser

Differential Revision: [D40869846](https://our.internmc.facebook.com/intern/diff/D40869846)
Pull Request resolved: pytorch#87779
Approved by: https://github.com/davidberard98
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@jjsjann123 jjsjann123 jjsjann123 left review comments

@csarofeen csarofeen csarofeen approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@shmsong @IvanYashchuk @csarofeen @jjsjann123