{numlib}[GCC/11.2.0] OpenBLAS v0.3.17 #13580

SebastianAchilles · 2021-08-03T22:40:36Z

(created using eb --new-pr)

Micket

LGTM

Micket · 2021-08-03T23:36:23Z

@boegelbot please test @ generoso

Micket · 2021-08-03T23:37:59Z

Test report by @Micket
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
alvis-c1 - Linux centos linux 7.9.2009, x86_64, Intel Xeon Processor (Skylake), Python 3.6.8
See https://gist.github.com/b482c399f56ccbbdf775ea688bb44254 for a full test report.

boegelbot · 2021-08-03T23:40:09Z

@Micket: Request for testing this PR well received on generoso

PR test command 'EB_PR=13580 EB_ARGS= /apps/slurm/default/bin/sbatch --job-name test_PR_13580 --ntasks=4 ~/boegelbot/eb_from_pr_upload_generoso.sh' executed!

exit code: 0
output:

Submitted batch job 17879

Test results coming soon (I hope)...

- notification for comment with ID 892236029 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).

Micket · 2021-08-03T23:43:54Z

oh-oh, i got precision errors

boegelbot · 2021-08-03T23:45:17Z

Test report by @boegelbot
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
generoso-c1-s-4 - Linux centos linux 8.2.2004, x86_64, Intel(R) Xeon(R) CPU E5-2667 v3 @ 3.20GHz (haswell), Python 3.6.8
See https://gist.github.com/a9771dfda7a0bc05dfd5f56748c6056d for a full test report.

branfosj · 2021-08-04T07:41:16Z

oh-oh, i got precision errors

I also see these on my Intel systems. There tests pass on POWER9. We've already built OpenBLAS 0.3.17 (#13421).

SebastianAchilles · 2021-08-04T09:25:05Z

Test report by @SebastianAchilles
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
rocky8-eb - Linux rocky linux 8.4, x86_64, Intel(R) Core(TM) i7-6900K CPU @ 3.20GHz (broadwell), Python 3.6.8
See https://gist.github.com/47afbcfddb256aa3ecc9d35e556b93a8 for a full test report.

Micket · 2021-08-04T09:25:22Z

Test report by @Micket
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
alvis-c1 - Linux centos linux 7.9.2009, x86_64, Intel Xeon Processor (Skylake), Python 3.6.8
See https://gist.github.com/7a241491e7e6dd9047d255dead558d84 for a full test report.

SebastianAchilles · 2021-08-04T09:27:07Z

Test report by @SebastianAchilles
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
centos8-eb - Linux centos linux 8.3.2011, x86_64, Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz (skylake), Python 3.6.8
See https://gist.github.com/c612f55c49267ab422bf973733be45bc for a full test report.

boegel · 2021-08-04T09:40:56Z

@boegelbot please test @ generoso

boegel · 2021-08-04T09:42:23Z

Looks like disabling the use of -free-vectorize helps... Does that ring a bell for you @bartoldeman? ;)

@SebastianAchilles Maybe add a comment in the easyconfig to explain why we disable vectorize?
Should we open an issue upstream for this?

boegelbot · 2021-08-04T09:45:11Z

@boegel: Request for testing this PR well received on generoso

PR test command 'EB_PR=13580 EB_ARGS= /apps/slurm/default/bin/sbatch --job-name test_PR_13580 --ntasks=4 ~/boegelbot/eb_from_pr_upload_generoso.sh' executed!

exit code: 0
output:

Submitted batch job 17885

Test results coming soon (I hope)...

- notification for comment with ID 892516697 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).

boegel · 2021-08-04T09:45:33Z

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
node3542.doduo.os - Linux RHEL 8.2, x86_64, AMD EPYC 7552 48-Core Processor (zen2), Python 3.6.8
See https://gist.github.com/c3c84cef76e38dd359784839422d3ae1 for a full test report.

Micket · 2021-08-04T09:50:16Z

For reference, with vectorization on, I see

 ******* FATAL ERROR - COMPUTED RESULT IS LESS THAN HALF ACCURATE *******
                       EXPECTED RESULT                    COMPUTED RESULT
       1  (   0.290119E-01,   0.379002E-01)  (  -0.408423E-01,  -0.319541E-01)
       2  (   0.762275E-02,   0.806386E-02)  (  -0.924949E-02,  -0.880837E-02)
 ******* ZGEMM  FAILED ON CALL NUMBER:
   6457: ZGEMM ('C','N',  2,  1,  1,( 1.0, 0.0), A,  2, B,  2,( 0.0, 0.0), C,  3).

 ZHEMM  PASSED THE COMPUTATIONAL TESTS (  1296 CALLS)

 ZSYMM  PASSED THE COMPUTATIONAL TESTS (  1296 CALLS)

 ******* FATAL ERROR - COMPUTED RESULT IS LESS THAN HALF ACCURATE *******
                       EXPECTED RESULT                    COMPUTED RESULT
       1  (   0.306693    ,   0.242757    )  (   0.306693    ,   0.242757    )
       2  (   0.360524    ,   0.364897    )  (   0.379238    ,   0.383611    )
 ******* ZTRMM  FAILED ON CALL NUMBER:
    950: ZTRMM ('L','U','C','U',  2,  1,( 1.0, 0.0), A,  3, B,  3)               .

 ******* FATAL ERROR - COMPUTED RESULT IS LESS THAN HALF ACCURATE *******
                       EXPECTED RESULT                    COMPUTED RESULT
       1  (   0.306693    ,   0.242757    )  (   0.306693    ,   0.242757    )
       2  (   0.466533    ,   0.306693    )  (   0.447819    ,   0.287979    )
 ******* ZTRSM  FAILED ON CALL NUMBER:
    950: ZTRSM ('L','U','C','U',  2,  1,( 1.0, 0.0), A,  3, B,  3)               .

 ******* FATAL ERROR - COMPUTED RESULT IS LESS THAN HALF ACCURATE *******
                       EXPECTED RESULT                    COMPUTED RESULT
       1  (   0.387794    ,   0.256287E-02)  (   0.246317    ,  -0.138914    )
       2  (   0.227563    ,   0.341716E-01)  (   0.193391    ,    0.00000    )
      THESE ARE THE RESULTS FOR COLUMN   2
 ******* ZHERK  FAILED ON CALL NUMBER:
    490: ZHERK ('U','C',  2,  1, 1.0, A,  2, 0.0, C,  3)                         .

 ZSYRK  PASSED THE COMPUTATIONAL TESTS (  1296 CALLS)

 ******* FATAL ERROR - COMPUTED RESULT IS LESS THAN HALF ACCURATE *******
                       EXPECTED RESULT                    COMPUTED RESULT
       1  (   0.131519    ,  -0.747125E-01)  (   0.182635E-02,  -0.204405    )
       2  (   0.953971E-01,   0.193181    )  (  -0.977843E-01,    0.00000    )
      THESE ARE THE RESULTS FOR COLUMN   2
 ******* ZHER2K FAILED ON CALL NUMBER:
    490: ZHER2K('U','C',  2,  1,( 1.0, 0.0), A,  2, B,  2, 0.0, C,  3)           .

 ******* FATAL ERROR - COMPUTED RESULT IS LESS THAN HALF ACCURATE *******
                       EXPECTED RESULT                    COMPUTED RESULT
       1  (   0.495928E-01,   0.197144    )  (  -0.176908    ,  -0.293573E-01)
       2  (  -0.154221    ,  -0.169237    )  (   0.179134    ,   0.164118    )
 ******* cblas_zgemm  FAILED ON CALL NUMBER:
   8644: cblas_zgemm ( CblasColMajor,CblasConjTrans,  CblasNoTrans,
            2,  1,  1, ( 1.0, 0.0) , A,  2, B,  2, ( 0.0, 0.0) , C,  3).
 ******* cblas_zgemm  FAILED ON CALL NUMBER:
   4537: cblas_zgemm ( CblasRowMajor,  CblasNoTrans,  CblasNoTrans,
            1,  1,  0, ( 0.0, 0.0) , A,  2, B,  1, ( 0.0, 0.0) , C,  2).

boegelbot · 2021-08-04T09:50:17Z

Test report by @boegelbot
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
generoso-c1-s-5 - Linux centos linux 8.2.2004, x86_64, Intel(R) Xeon(R) CPU E5-2667 v3 @ 3.20GHz (haswell), Python 3.6.8
See https://gist.github.com/e096206993e6f443ec1a4adc319c16e6 for a full test report.

boegel · 2021-08-04T10:08:37Z

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
node3107.skitty.os - Linux centos linux 7.9.2009, x86_64, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz, Python 3.6.8
See https://gist.github.com/bdafec866fb628dece3bb6bce4cff176 for a full test report.

boegel · 2021-08-04T11:08:34Z

Test report by @boegel
SUCCESS
Build succeeded for 11 out of 11 (1 easyconfigs in total)
node2609.swalot.os - Linux centos linux 7.9.2009, x86_64, Intel(R) Xeon(R) CPU E5-2660 v3 @ 2.60GHz (haswell), Python 3.6.8
See https://gist.github.com/4b9f177b0de43c320f5b5c010fc9705c for a full test report.

…lain why we disable vectorize

boegel · 2021-08-04T11:42:05Z

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
select-pika-c6gd-2xlarge-0001 - Linux centos linux 8.3.2011, AArch64, ARM UNKNOWN (graviton2), Python 3.6.8
See https://gist.github.com/cf62ba3e2375f3a24b951ea0dee6eb8c for a full test report.

boegel · 2021-08-04T14:16:00Z

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in total)
easybuild2.novalocal - Linux centos linux 8.3.2011, POWER, IBM pSeries (emulated by qemu) (power9le), Python 3.6.8
See https://gist.github.com/b4578ce3c743e8a0f84dd7d97768347c for a full test report.

boegel

lgtm

boegel · 2021-08-04T18:58:27Z

Going in, thanks @SebastianAchilles!

adding easyconfigs: OpenBLAS-0.3.17-GCC-11.2.0.eb

7ec2ef1

SebastianAchilles added the update label Aug 3, 2021

SebastianAchilles added this to the 4.x milestone Aug 3, 2021

Micket previously approved these changes Aug 3, 2021

View reviewed changes

update easyconfig OpenBLAS-0.3.17-GCC-11.2.0.eb: disable vectorize

d07d9ac

SebastianAchilles dismissed Micket’s stale review via d07d9ac August 4, 2021 09:20

boegel modified the milestones: 4.x, next release (4.4.2?) Aug 4, 2021

update easyconfig OpenBLAS-0.3.17-GCC-11.2.0.eb: add a comment to exp…

8edfb42

…lain why we disable vectorize

boegel approved these changes Aug 4, 2021

View reviewed changes

boegel merged commit d7620e8 into easybuilders:develop Aug 4, 2021

SebastianAchilles mentioned this pull request Aug 5, 2021

add patch for OpenBLAS v0.3.17 to workaround problem in C/ZBLAS3 tests with GCC 11.x and -ftree-vectorize #13625

Merged

boegel mentioned this pull request Feb 12, 2024

replace run_cmd with run_shell_cmd in custom easyblock for OpenBLAS (openblas.py) easybuilders/easybuild-easyblocks#3195

Merged

{numlib}[GCC/11.2.0] OpenBLAS v0.3.17 #13580

{numlib}[GCC/11.2.0] OpenBLAS v0.3.17 #13580

Uh oh!

Conversation

SebastianAchilles commented Aug 3, 2021

Uh oh!

Micket left a comment

Choose a reason for hiding this comment

Uh oh!

Micket commented Aug 3, 2021

Uh oh!

Micket commented Aug 3, 2021

Uh oh!

boegelbot commented Aug 3, 2021

Uh oh!

Micket commented Aug 3, 2021

Uh oh!

boegelbot commented Aug 3, 2021

Uh oh!

branfosj commented Aug 4, 2021

Uh oh!

SebastianAchilles commented Aug 4, 2021

Uh oh!

Micket commented Aug 4, 2021

Uh oh!

SebastianAchilles commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

boegelbot commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

Micket commented Aug 4, 2021

Uh oh!

boegelbot commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

boegel left a comment

Choose a reason for hiding this comment

Uh oh!

boegel commented Aug 4, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants