{chem}[foss/2023a] VeloxChem 1.0-rc4 by Crivella · Pull Request #24176 · easybuilders/easybuild-easyconfigs

Crivella · 2025-10-09T14:00:07Z

Add EC file for VeloxChem

Crivella · 2025-10-09T14:03:53Z

@boegelbot please test @ jsc-zen3

boegelbot · 2025-10-09T14:10:12Z

@Crivella: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=24176 EB_ARGS= EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_24176 --ntasks=8 ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 8247

Test results coming soon (I hope)...

Details

- notification for comment with ID 3386032242 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

boegelbot · 2025-10-09T15:22:42Z

Test report by @boegelbot
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
jsczen3c1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC-Milan Processor (zen3), Python 3.9.21
See https://gist.github.com/boegelbot/b22cc7f23e255d81351bc85e4658a880 for a full test report.

boegel · 2025-10-13T11:25:22Z

@Crivella Does this depend on some other PR/fix?

When I test this, I get a failure:

  -- Setting option VLX_LA_VENDOR: FlexiBLAS
  CMake Error at cmake/custom/math.cmake:278 (message):
    Cannot find a working CBLAS library to link to!

boegel · 2025-10-13T11:26:00Z

Test report by @boegel
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
node3512.doduo.os - Linux RHEL 9.6, x86_64, AMD EPYC 7552 48-Core Processor (zen2), Python 3.9.21
See https://gist.github.com/boegel/40b141f2ac11b2dc46b9f5be03ba3238 for a full test report.

Crivella · 2025-10-13T11:29:25Z

Not sure if it is related to

add support for iterative build of OpenBLAS with 64-bit integers easybuild-easyblocks#2753

or something similar to

Explicitly set LAPACK/BLAS to FlexiBLAS if present in OpenCV easybuild-easyblocks#3874

but i think having a recent rebuild of FlexiBLAS might help.

I would assume it is something already in develop since the bot succeeded?

Crivella · 2025-10-13T11:32:32Z

Worst case we can set it to OpenBLAS the same way they do in

https://gitlab.com/veloxchem/veloxchem/-/blob/v1.0-rc4/.github/workflows/ci.yml#L73

Crivella · 2025-10-13T11:41:55Z

NOTE this is what i get

  -- Setting (unspecified) option VLX_ALIGN: 64
  -- Setting option VLX_LA_VENDOR: FlexiBLAS
  -- Using FlexiBLAS as linear algebra backend
  -- BLAS/LAPACK include directories: /home/crivella/.local/easybuild/software/OpenBLAS/0.3.23-GCC-12.3.0/include
  -- BLAS/LAPACK libraries: /home/crivella/.local/easybuild/software/FlexiBLAS/3.3.1-GCC-12.3.0/lib/liblapacke.so;/home/crivella/.local/easybuild/software/FlexiBLAS/3.3.1-GCC-12.3.0/lib/libflexiblas.so

EDIT: which on second inspection might need some extra sauce to avoid the same problem as in

Explicitly set LAPACK/BLAS to FlexiBLAS if present in OpenCV easybuild-easyblocks#3874

of mixing FlexiBLAS libs with the OpenBLAS headers

boegel · 2025-10-13T12:22:07Z

@Crivella Why doesn't this happen consistently though?

Crivella · 2025-10-13T12:55:31Z

The CMake part for the lapack/blas seems to rely either on LAPACK_ROOT/BLAS_ROOT or on known paths in CMAKE_PREFIX_PATH

https://gitlab.com/veloxchem/veloxchem/-/blob/v1.0-rc4/cmake/custom/math.cmake?ref_type=tags#L226-235

I am trying to fiddle with the first to force it to pick all stuff from FlexiBLAS.

Concerning the failure in your build i am not sure why it is not finding cblas.h is it possible CMAKE_PREFIX_PATH is not being set somehow?

Crivella · 2025-10-13T14:16:06Z

Another note, using FlexiBLAS or any string that is not recognized is defering to the find_path section.

...
elseif(VLX_LA_VENDOR STREQUAL "Generic")
  find_package(LAPACK REQUIRED QUIET)

  set(_check_cblas_and_lapacke TRUE)
else()
  find_package(LAPACK REQUIRED QUIET)

  set(_check_cblas_and_lapacke TRUE)
endif()

The only special values for VLX_LA_VENDOR from the cmake file are

MKL
Apple or NAS
Cray
FLAME

Crivella · 2025-10-13T14:34:14Z

@boegelbot please test @ jsc-zen3

boegelbot · 2025-10-13T14:40:13Z

@Crivella: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=24176 EB_ARGS= EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_24176 --ntasks=8 ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 8334

Test results coming soon (I hope)...

Details

- notification for comment with ID 3397788101 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

Crivella · 2025-10-13T14:41:28Z

Test report by @Crivella
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
crivella-desktop - Linux Ubuntu 22.04.5 LTS (Jammy Jellyfish), x86_64, 13th Gen Intel(R) Core(TM) i9-13900K (skylake), Python 3.11.13
See https://gist.github.com/Crivella/61ac75ceaffdd7e8a4a5f849b5613927 for a full test report.

NOTES

  -- Setting option VLX_LA_VENDOR: FlexiBLAS
  -- Using FlexiBLAS as linear algebra backend
  -- BLAS/LAPACK include directories: /home/crivella/.local/easybuild/software/FlexiBLAS/3.3.1-GCC-12.3.0/include/flexiblas
  -- BLAS/LAPACK libraries: /home/crivella/.local/easybuild/software/FlexiBLAS/3.3.1-GCC-12.3.0/lib/liblapacke.so;/home/crivella/.local/easybuild/software/FlexiBLAS/3.3.1-GCC-12.3.0/lib/libflexiblas.so
  -- Searching for Eigen in /home/crivella/.local/easybuild/software/Eigen/3.4.0-GCCcore-12.3.0/include/Eigen
  -- Found Eigen in /home/crivella/.local/easybuild/software/Eigen/3.4.0-GCCcore-12.3.0/include/Eigen

boegelbot · 2025-10-13T15:52:09Z

Test report by @boegelbot
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
jsczen3c1.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC-Milan Processor (zen3), Python 3.9.21
See https://gist.github.com/boegelbot/afecd9df8de4f5d2da14831824a0ba5f for a full test report.

boegel · 2025-10-14T18:00:52Z

Test report by @boegel
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
node3596.doduo.os - Linux RHEL 9.6, x86_64, AMD EPYC 7552 48-Core Processor (zen2), Python 3.9.21
See https://gist.github.com/boegel/50a8424dd77e2149b3bbba08999171a1 for a full test report.

Crivella · 2025-10-15T08:01:27Z

Now it is picking up the libraries but failing somewhere else.

Another difference i noticed in the logs, is for your build, the test for using the Ninja generator is failing and falling back to Unix makefiles in scikit-build.
Would be nice to see from logs in the bot build if that is also the case or if picking up Ninja is working there.

I would assume it is easier to solve this problem than whatever is causing the Makefiles build version to fail

From my logs:

  --------------------------------------------------------------------------------
  -- Trying 'Ninja' generator
  --------------------------------
  ---------------------------
  ----------------------
  -----------------
  ------------
  -------
  --
  Not searching for unused variables given on the command line.
  -- The C compiler identification is GNU 12.3.0
  -- Detecting C compiler ABI info
  -- Detecting C compiler ABI info - done
  -- Check for working C compiler: /tmp/eb-8k0nmnof/tmp2v_vn4yf/rpath_wrappers/gcc_wrapper/gcc - skipped
  -- Detecting C compile features
  -- Detecting C compile features - done
  -- The CXX compiler identification is GNU 12.3.0
  -- Detecting CXX compiler ABI info
  -- Detecting CXX compiler ABI info - done
  -- Check for working CXX compiler: /tmp/eb-8k0nmnof/tmp2v_vn4yf/rpath_wrappers/gxx_wrapper/g++ - skipped
  -- Detecting CXX compile features
  -- Detecting CXX compile features - done
  -- Configuring done (2.1s)
  -- Generating done (0.0s)
  -- Build files have been written to: /home/crivella/.local/easybuild/build/VeloxChem/1.0-rc4/foss-2023a/VeloxChem/veloxchem-v1.0-rc4-770fddbeda827eb483a1919689e0d289e72a2c5c/_cmake_test_compile/build
  --
  -------
  ------------
  -----------------
  ----------------------
  ---------------------------
  --------------------------------
  -- Trying 'Ninja' generator - success
  --------------------------------------------------------------------------------

Crivella · 2025-10-15T13:19:44Z

The reason this is failing is the other way around (will fail with any new build of OpenBLAS after #2753) i thought i had a newer build of OpenBLAS/FlexiBLAS but not for this TC (and i guess the bot as well).

The error in the frist failed report was due to

  check_cblas_lapacke(_la_cblas_bundled
    CBLAS
    cblas_sgemm
    ${_include_dirs}
    "cblas.h"
    ""
    "${_blas_libs}"
    ""
    "${BLAS_ROOT};${LAPACK_ROOT}"
    "lib"
    )

failing the check (i assume since cblas_sgemm64_ is what is defined in cblas.h )

With the patch now it is picking up the include and lib folders from FlexiBLAS so it is passing that check, but for some reason during the compilation it still seems to try and use the file from OpenBLAS
'cblas_dgemm' was not declared in this scope; did you mean 'cblas_dgemm64_'

I will try to recompile a newer OpenBLAS/FlexiBLAS to reproduce the error and see how it can be fixed

EDIT

VeloxChem build system passes the include dir from the detected cblas.h as a system header with -isystem .
While we have FlexiBLAS also in CPATH, before OpenBLAS, g++ must be doing some filtering so that the CLI -isystem takes precedence and apparently -isystem locations are searched after CPATH ones

Crivella · 2025-10-15T16:04:14Z

@boegelbot please test @ jsc-zen3
CORE_CNT=16
EB_ARGS="--installpath /tmp/$USER/pr-24176 OpenBLAS-0.3.23-GCC-12.3.0.eb FlexiBLAS-3.3.1-GCC-12.3.0.eb VeloxChem-1.0-rc4-foss-2023a.eb"

boegelbot · 2025-10-15T16:10:10Z

@Crivella: Request for testing this PR well received on jsczen3l1.int.jsc-zen3.fz-juelich.de

PR test command 'if [[ develop != 'develop' ]]; then EB_BRANCH=develop ./easybuild_develop.sh 2> /dev/null 1>&2; EB_PREFIX=/home/boegelbot/easybuild/develop source init_env_easybuild_develop.sh; fi; EB_PR=24176 EB_ARGS="--installpath /tmp/$USER/pr-24176 OpenBLAS-0.3.23-GCC-12.3.0.eb FlexiBLAS-3.3.1-GCC-12.3.0.eb VeloxChem-1.0-rc4-foss-2023a.eb" EB_CONTAINER= EB_REPO=easybuild-easyconfigs EB_BRANCH=develop /opt/software/slurm/bin/sbatch --job-name test_PR_24176 --ntasks="16" ~/boegelbot/eb_from_pr_upload_jsc-zen3.sh' executed!

exit code: 0
output:

Submitted batch job 8412

Test results coming soon (I hope)...

Details

- notification for comment with ID 3407215490 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

Crivella · 2025-10-15T16:23:53Z

Test report by @Crivella
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
crivella-desktop - Linux Ubuntu 22.04.5 LTS (Jammy Jellyfish), x86_64, 13th Gen Intel(R) Core(TM) i9-13900K (skylake), Python 3.11.13
See https://gist.github.com/Crivella/08797f6cd8b6408778e1582427d9a8fb for a full test report.

boegelbot · 2025-10-15T16:56:28Z

Test report by @boegelbot
SUCCESS
Build succeeded for 3 out of 3 (3 easyconfigs in total)
jsczen3c2.int.jsc-zen3.fz-juelich.de - Linux Rocky Linux 9.6, x86_64, AMD EPYC-Milan Processor (zen3), Python 3.9.21
See https://gist.github.com/boegelbot/7b271f739c5057b658fcdc4891c18f03 for a full test report.

boegel · 2025-10-21T07:39:15Z

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
node3541.doduo.os - Linux RHEL 9.6, x86_64, AMD EPYC 7552 48-Core Processor (zen2), Python 3.9.21
See https://gist.github.com/boegel/d7f9fbbd9c681755e16fb763f12a279e for a full test report.

…in VeloxChem easyconfig

boegel · 2025-10-21T11:02:53Z

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
node4206.shinx.os - Linux RHEL 9.6, x86_64, AMD EPYC 9654 96-Core Processor (zen4), Python 3.9.21
See https://gist.github.com/boegel/3c92ffa36212b1a993f324bdc3e2e718 for a full test report.

boegel

lgtm

boegel · 2025-10-21T14:56:41Z

Going in, thanks @Crivella!

Add EC file for VeloxChem

103cb87

Crivella added the 2023a label Oct 9, 2025

github-actions Bot added the new label Oct 9, 2025

Downgrade pybind

c46d8d0

Crivella added the ready-to-review label Oct 9, 2025

Crivella added this to the next release (5.2.0?) milestone Oct 9, 2025

boegel self-assigned this Oct 13, 2025

Add patch to allow FlexiBLAS's cblas.h to be used

a88269e

Crivella added 2 commits October 15, 2025 16:13

Allow scikit to build with ninja

01668f9

Ensure FlexiBLAS cblas.h can be used

715d583

Crivella removed the ready-to-review label Oct 15, 2025

Crivella added the ready-to-review label Oct 16, 2025

Crivella requested a review from boegel October 16, 2025 09:35

boegel requested changes Oct 21, 2025

View reviewed changes

Comment thread easybuild/easyconfigs/v/VeloxChem/VeloxChem-1.0-rc4-foss-2023a.eb Outdated

add more specific sanity check path for Python site-packages lib dir …

5043e56

…in VeloxChem easyconfig

boegel approved these changes Oct 21, 2025

View reviewed changes

boegel merged commit 107ea0e into easybuilders:develop Oct 21, 2025
8 checks passed

Crivella deleted the feature-VeloxChem branch October 21, 2025 15:49

Conversation

Crivella commented Oct 9, 2025

Uh oh!

Crivella commented Oct 9, 2025

Uh oh!

boegelbot commented Oct 9, 2025

Uh oh!

boegelbot commented Oct 9, 2025

Uh oh!

boegel commented Oct 13, 2025

Uh oh!

boegel commented Oct 13, 2025

Uh oh!

Crivella commented Oct 13, 2025

Uh oh!

Crivella commented Oct 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Crivella commented Oct 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

boegel commented Oct 13, 2025

Uh oh!

Crivella commented Oct 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Crivella commented Oct 13, 2025

Uh oh!

Crivella commented Oct 13, 2025

Uh oh!

boegelbot commented Oct 13, 2025

Uh oh!

Crivella commented Oct 13, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

NOTES

Uh oh!

boegelbot commented Oct 13, 2025

Uh oh!

boegel commented Oct 14, 2025

Uh oh!

Crivella commented Oct 15, 2025

Uh oh!

Crivella commented Oct 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Crivella commented Oct 15, 2025

Uh oh!

boegelbot commented Oct 15, 2025

Uh oh!

Crivella commented Oct 15, 2025

Uh oh!

boegelbot commented Oct 15, 2025

Uh oh!

boegel commented Oct 21, 2025

Uh oh!

Uh oh!

boegel commented Oct 21, 2025

Uh oh!

boegel left a comment

Choose a reason for hiding this comment

Uh oh!

boegel commented Oct 21, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Crivella commented Oct 13, 2025 •

edited

Loading

Crivella commented Oct 13, 2025 •

edited

Loading

Crivella commented Oct 13, 2025 •

edited

Loading

Crivella commented Oct 13, 2025 •

edited

Loading

Crivella commented Oct 15, 2025 •

edited

Loading