8000 BLD: Fix x86-simd-sort build failure on openBSD by charris · Pull Request #26821 · numpy/numpy · GitHub
[go: up one dir, main page]
Skip to content

BLD: Fix x86-simd-sort build failure on openBSD #26821

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
Merged
merged 5 commits into from
Jul 4, 2024
Merged

BLD: Fix x86-simd-sort build failure on openBSD #26821

merged 5 commits into from
Jul 4, 2024

Conversation

charris
Copy link
Member
@charris charris commented Jul 1, 2024 •
edited
Loading

Backport of #26797.

Fixes building x86-simd-sort on openBSD. See intel/x86-simd-sort#157 for details.

The following are updated from main

  • highway submodule
  • vendored-meson submodule
  • x86_simd_sort submodule
  • highway-qsort.dispatch.cpp.
  • x86_simd_argsort.dispatch.cpp
  • x86_simd_qsort.dispatch.cpp
  • x86_simd_qsort_16bit.dispatch.cpp

Small fixes are taken from

  • numpy/_core/meson.build

@charris charris added 00 - Bug 03 - Maintenance 08 - Backport Used to tag backport PRs 36 - Build Build related PR labels Jul 1, 2024
@charris charris added this to the 2.0.1 release milestone Jul 1, 2024
@charris
Copy link
Member Author
charris commented Jul 1, 2024

@rgommers Should we update vendored-meson as well. I've been conservative about that.

@charris
Copy link
Member Author
charris commented Jul 1, 2024 •
edited
Loading

@r-devulap I assume that there is a missing backport here. Do you know what it is?

EDIT: I may just omit this backport, it was a maybe yes, maybe no decision to try.

@r-devulap
Copy link
Member
r-devulap commented Jul 1, 2024 •
edited
Loading

probably this one? #26338

@rgommers
Copy link
Member
rgommers commented Jul 1, 2024

@rgommers Should we update vendored-meson as well. I've been conservative about that.

Working on it at numpy/meson#15

@charris
Copy link
Member Author
charris commented Jul 1, 2024

Working on it at

I suspect the new risc failure may be meson connected.

r-devulap and others added 4 commits July 2, 2024 12:49
@r-devulap @charris
BLD: Fix x86-simd-sort build failure on openBSD
8d9b766
@charris
MAINT: Update highway to 1dbb1180.
2ccc50a
@charris
MAINT: Update various sort files from main
cd51f2d 
- Update highway-qsort.dispatch.cpp.
- Update x86_simd_argsort.dispatch.cpp
- Update x86_simd_qsort.dispatch.cpp
- Update x86_simd_qsort_16bit.dispatch.cpp
@charris
MAINT: Update vendored-meson from main
8ed0fb0
@mattip
Copy link
Member
mattip commented Jul 3, 2024

There is a failure on the riscv qemu build @r-devulap does this ring a bell?

FAILED: numpy/_core/libhighway.a.p/src_highway_hwy_targets.cc.o 
c++ -Inumpy/_core/libhighway.a.p -Inumpy/_core -I../numpy/_core -I../numpy/_core/src/highway -I/numpy/build/meson_cpu -fdiagnostics-color=always -D_GLIBCXX_ASSERTIONS=1 -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -std=c++17 -O2 -g -fPIC -DTOOLCHAIN_MISS_ASM_HWCAP_H -MD -MQ numpy/_core/libhighway.a.p/src_highway_hwy_targets.cc.o -MF numpy/_core/libhighway.a.p/src_highway_hwy_targets.cc.o.d -o numpy/_core/libhighway.a.p/src_highway_hwy_targets.cc.o -c ../numpy/_core/src/highway/hwy/targets.cc
../numpy/_core/src/highway/hwy/targets.cc: Assembler messages:
../numpy/_core/src/highway/hwy/targets.cc:631: Error: unrecognized opcode `vsetvli s0,zero,e8,m1,ta,ma'

@r-devulap
Copy link
Member
r-devulap commented Jul 3, 2024 •
edited
Loading

@mattip I think you need to backport this PR #26273 as well. The changes in numpy/_core/meson.build seem relevant.

@charris
Copy link
Member Author
charris commented Jul 3, 2024

I think you need to backport this PR #26273 as well.

Looking at the diff of numpy/_core/meson.build, there seem to be a few other missing pieces as well. Is the problem fixed by #26797 also present in 2.0.0? If not, I'm inclined to skip this backport, it is getting overly intrusive.

@r-devulap
Copy link
Member
r-devulap commented Jul 4, 2024 •
edited
Loading

Looking at the diff of numpy/_core/meson.build, there seem to be a few other missing pieces as well. Is the problem fixed by #26797 also present in 2.0.0? If not, I'm inclined to skip this backport, it is getting overly intrusive.

unfortunately, yes :( The original bug report was for numpy 2.0.0

@charris charris merged commit ba7dfa0 into numpy:maintenance/2.0.x Jul 4, 2024
58 of 59 checks passed
@charris charris deleted the backport-26797 branch July 4, 2024 20:23
This was referenced Jul 25, 2024
Closed
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
00 - Bug 03 - Maintenance 08 - Backport Used to tag backport PRs 36 - Build Build related PR
Projects
None yet
Milestone
2.0.1 release
Development

Successfully merging this pull request may close these issues.
4 participants
@charris @r-devulap @rgommers @mattip
0