Help with flexiblas crash on aarch64 in kojij only

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Since the latest update to OpenBLAS 0.3.28 in rawhide, FlexiBLAS fails to build in aarch64 because OpenBLAS crashes in the LAPACK-xeigtstc_cec_in test. Note that OpenBLAS itself does not fail only because they don't include LAPACK test suite.

See:
- The first failure in Koschei after the 0.3.28 update: https://koschei.fedoraproject.org/package/flexiblas - The build log: https://koji.fedoraproject.org/koji/taskinfo?taskID=125998498

FTBFS report here: https://bugzilla.redhat.com/show_bug.cgi?id=2329491

I have attempted to collect some more debug info via the following - https://src.fedoraproject.org/fork/orion/rpms/flexiblas/tree/debug

But the valgrind run just seems to hang with no output from valgrind - https://kojipkgs.fedoraproject.org//work/tasks/3875/127513875/build.log

 Tests of the Nonsymmetric eigenproblem condition estimation routines
 CTRSYL, CTREXC, CTRSNA, CTRSEN
 Relative machine precision (EPS) =     0.119209E-06
 Safe minimum (SFMIN)             =     0.117549E-37
 Routines pass computational tests if test ratio is less than   20.00
 CEC routines passed the tests of the error exits ( 41 tests done)

And the crash seems to occur after memory corruption has already occurred so seems to be of limited utility. So I'm at a loss myself.

It appears to build fine in COPR and on the Fedora dev test machine aarch64-test01.fedorainfracloud.org so we seem to be stuck with the limited access koji builders. I guess this points to possible architectural differences, but cpuinfo doesn't show anything:

aarch-test01:
Features : fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp
CPU implementer : 0x41
CPU architecture: 8
CPU variant     : 0x3
CPU part        : 0xd0c
CPU revision    : 1

koji:
Features : fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp
CPU implementer	: 0x41
CPU architecture: 8
CPU variant	: 0x3
CPU part	: 0xd0c
CPU revision	: 1

Koji does have 12 cores as opposed to 2, so it might be due to how openblas splits up tasks among the cores - but not sure why it's only showing up on aarch64 unless 12 cores is a magic number.

Any other suggestions?

--
Orion Poplawski
he/him/his  - surely the least important thing about me
IT Systems Manager                         720-772-5637
NWRA, Boulder/CoRA Office             FAX: 303-415-9702
3380 Mitchell Lane                       orion@xxxxxxxx
Boulder, CO 80301                 https://www.nwra.com/

<<attachment: smime.p7s>>

-- 
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Users]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]

  Powered by Linux