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

New LLVM vectorizer crash in assertion #91883

Closed
cheshire opened this issue May 12, 2024 · 0 comments
Closed

New LLVM vectorizer crash in assertion #91883

cheshire opened this issue May 12, 2024 · 0 comments
Labels
crash Prefer [crash-on-valid] or [crash-on-invalid] vectorization

Comments

@cheshire
Copy link
Contributor

@fhahn hits an assertion introduced in #80271

Input file: https://gist.github.com/cheshire/e504c63db9bdb300fbdcc1df2e183bd8

Repro:

opt -S -passes=loop-vectorize -mattr=+prfchw,-cldemote,+avx,+aes,+sahf,+pclmul,-xop,+crc32,+xsaves,-avx512fp16,-usermsr,-sm4,-egpr,+sse4.1,-avx512ifma,+xsave,+sse4.2,-avx512pf,-tsxldtrk,-ptwrite,-widekl,-sm3,+invpcid,+64bit,+xsavec,-avx10.1-512,-avx512vpopcntdq,+cmov,-avx512vp2intersect,+avx512cd,+movbe,-avxvnniint8,-avx512er,-ccmp,-amx-int8,-kl,-avx10.1-256,+evex512,-avxvnni,+rtm,+adx,+avx2,-hreset,-movdiri,-serialize,-sha512,-vpclmulqdq,+avx512vl,-uintr,-cf,+clflushopt,-raoint,-cmpccxadd,+bmi,-amx-tile,+sse,-gfni,-avxvnniint16,-amx-fp16,-ndd,+xsaveopt,+rdrnd,+avx512f,-amx-bf16,-avx512bf16,-avx512vnni,-push2pop2,+cx8,+avx512bw,+sse3,+pku,+fsgsbase,-clzero,-mwaitx,-lwp,+lzcnt,-sha,-movdir64b,-ppx,-wbnoinvd,-enqcmd,-prefetchwt1,-avxneconvert,-tbm,-pconfig,-amx-complex,+ssse3,+cx16,+bmi2,+fma,+popcnt,-avxifma,+f16c,-avx512bitalg,-rdpru,+clwb,+mmx,+sse2,+rdseed,-avx512vbmi2,-prefetchi,-rdpid,-fma4,-avx512vbmi,-shstk,-vaes,-waitpkg,-sgx,+fxsr,+avx512dq,-sse4a -mcpu=skylake-avx512 /tmp/minimized_crash.ll
@EugeneZelenko EugeneZelenko added vectorization crash Prefer [crash-on-valid] or [crash-on-invalid] and removed new issue labels May 12, 2024
@fhahn fhahn closed this as completed in 67d840b May 14, 2024
mub-at-arm pushed a commit to mub-at-arm/llvm-project that referenced this issue May 16, 2024
@fhahn @mub-at-arm
[VPlan] Relax over-aggressive assertion in VPTransformState::get().
f1f593c 
There are cases where a vector value has some users that demand the
the single scalar value only (NeedsScalar), while other users demand the
vector value (see attached test cases). In those cases, the NeedsScalar
users should only demand the first lane.

Fixes llvm#91883.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
crash Prefer [crash-on-valid] or [crash-on-invalid] vectorization
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@cheshire @EugeneZelenko