{toolchain} intel/2020b #11337

boegel · 2020-09-23T12:14:38Z

(created using eb --new-pr)

This is a candidate for intel/2020b...

boegel · 2020-09-23T12:54:11Z

@boegelbot please test @ generoso

boegelbot · 2020-09-23T13:05:11Z

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

PR test command 'EB_PR=11337 EB_ARGS= /apps/slurm/default/bin/sbatch --job-name test_PR_11337 ~/boegelbot/eb_from_pr_upload_generoso.sh' executed!

exit code: 0
output:

Submitted batch job 7877

Test results coming soon (I hope)...

- notification for comment with ID 697345114 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 · 2020-09-23T13:11:43Z

Test report by @boegelbot
FAILED
Build succeeded for 1 out of 6 (6 easyconfigs in this PR)
generoso-x-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/95e4fe185f22b9baf64958f52eaf4ae9 for a full test report.

edit (by @boegel): error setting up the boostrap proxies was caused by not having srun available via $PATH, see also #11425 (comment)

boegel · 2020-09-23T13:17:39Z

Test report by @boegel
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in this PR)
node2633.swalot.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) CPU E5-2660 v3 @ 2.60GHz (haswell), Python 2.7.5
See https://gist.github.com/d2bf8fcb45f6233ad636b1c6d017f099 for a full test report.

lexming · 2020-09-24T18:02:10Z

Test report by @lexming
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in this PR)
node128.hydra.os - Linux centos linux 7.7.1908, x86_64, Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz, Python 2.7.5
See https://gist.github.com/2b0051c80f8b9bfd47ce6e4f2ca315e4 for a full test report.

lexming · 2020-09-24T18:02:40Z

Test report by @lexming
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in this PR)
node376.hydra.os - Linux centos linux 7.7.1908, x86_64, Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz, Python 2.7.5
See https://gist.github.com/361d250b9c5df62ceb1b04cbc0f2f5f6 for a full test report.

easybuild/easyconfigs/i/impi/impi-2019.8.254-iccifort-2020.3.275.eb

…und impi bug, no longer relevant for impi 2019 update 8 Co-authored-by: Alex Domingo <alex.domingo.toro@vub.be>

boegel · 2020-10-05T18:18:56Z

@boegelbot please test @ generoso

boegelbot · 2020-10-05T18:20:08Z

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

PR test command 'EB_PR=11337 EB_ARGS= /apps/slurm/default/bin/sbatch --job-name test_PR_11337 ~/boegelbot/eb_from_pr_upload_generoso.sh' executed!

exit code: 0
output:

Submitted batch job 8031

Test results coming soon (I hope)...

- notification for comment with ID 703803892 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 · 2020-10-05T18:26:20Z

Test report by @boegelbot
FAILED
Build succeeded for 1 out of 6 (6 easyconfigs in this PR)
generoso-x-3 - 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/ad9264af3c8e83ca7e353947fb5b5e09 for a full test report.

edit (by @boegel):

[1601922345.426920] [generoso-x-3:3774850:0]         select.c:444  UCX  ERROR no active messages transport to <no debug data>: self/memory - Destination is unreachable
Abort(1091215) on node 1 (rank 1 in comm 0): Fatal error in PMPI_Init: Other MPI error, error stack:
MPIR_Init_thread(136)........: 
MPID_Init(1138)..............: 
MPIDI_OFI_mpi_init_hook(1541): OFI get address vector map failed

jhein32 · 2020-10-09T09:24:09Z

Hmm, doing a dry-run, it wants two versions of bison

-bash-4.2$ eb HPL-2.3-intel-2020.09.eb --robot --dry-run --use-existing-modules --from-pr=11337
== temporary log file in case of crash /tmp/eb-7Oh6BJ/easybuild-QdalMh.log
== found valid index for /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs, so using it...
Dry run: printing build status of easyconfigs and dependencies
 * [x] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/m/M4/M4-1.4.18.eb (module: Core | M4/1.4.18)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/b/Bison/Bison-3.7.1.eb (module: Core | Bison/3.7.1)
 * [x] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/b/Bison/Bison-3.3.2.eb (module: Core | Bison/3.3.2)
 * [x] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/z/zlib/zlib-1.2.11.eb (module: Core | zlib/1.2.11)
 * [x] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/h/help2man/help2man-1.47.4.eb (module: Core | help2man/1.47.4)
 * [x] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/f/flex/flex-2.6.4.eb (module: Core | flex/2.6.4)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/b/binutils/binutils-2.35.eb (module: Core | binutils/2.35)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/g/GCCcore/GCCcore-10.2.0.eb (module: Core | GCCcore/10.2.0)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/z/zlib/zlib-1.2.11-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | zlib/1.2.11)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/h/help2man/help2man-1.47.16-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | help2man/1.47.16)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/m/M4/M4-1.4.18-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | M4/1.4.18)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/b/Bison/Bison-3.7.1-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | Bison/3.7.1)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/f/flex/flex-2.6.4-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | flex/2.6.4)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/b/binutils/binutils-2.35-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | binutils/2.35)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/i/iccifort/iccifort-2020.3.275.eb (module: Core | iccifort/2020.3.275)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/p/pkg-config/pkg-config-0.29.2-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | pkg-config/0.29.2)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/l/libtool/libtool-2.4.6-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | libtool/2.4.6)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/e/expat/expat-2.2.9-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | expat/2.2.9)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/n/ncurses/ncurses-6.2-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | ncurses/6.2)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/l/libreadline/libreadline-8.0-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | libreadline/8.0)
 * [ ] /sw/easybuild/software/EasyBuild/4.3.0/easybuild/easyconfigs/p/Perl/Perl-5.32.0-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | Perl/5.32.0)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/a/Autoconf/Autoconf-2.69-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | Autoconf/2.69)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/a/Automake/Automake-1.16.2-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | Automake/1.16.2)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/a/Autotools/Autotools-20200321-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | Autotools/20200321)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/n/numactl/numactl-2.0.13-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | numactl/2.0.13)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/u/UCX/UCX-1.9.0-GCCcore-10.2.0.eb (module: Compiler/GCCcore/10.2.0 | UCX/1.9.0)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/i/impi/impi-2019.8.254-iccifort-2020.3.275.eb (module: Compiler/intel/2020.3.275 | impi/2019.8.254)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/i/iimpi/iimpi-2020.09.eb (module: Core | iimpi/2020.09)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/i/imkl/imkl-2020.3.279-iimpi-2020.09.eb (module: MPI/intel/2020.3.275/impi/2019.8.254 | imkl/2020.3.279)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/i/intel/intel-2020.09.eb (module: Core | intel/2020.09)
 * [ ] /tmp/eb-7Oh6BJ/files_pr11337/h/HPL/HPL-2.3-intel-2020.09.eb (module: MPI/intel/2020.3.275/impi/2019.8.254 | HPL/2.3)

boegel · 2020-10-09T09:30:16Z

Hmm, doing a dry-run, it wants two versions of bison

@jhein32 That's because of the bootstrapping mechanism that is done for binutils+GCCcore. We can consider bumping the Bison version that is used as an indirect build dep for the initial binutils, but that'll only fix the issue for recent toolchains.

jhein32 · 2020-10-12T09:01:38Z

I reported in #10899 that this set-up allows multi node running without any issues regarding UCX. Performance of the hpl is very good.

zao · 2020-10-16T16:35:56Z

Test report by @zao
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in total)
freja - Linux Ubuntu 20.04, x86_64, Intel(R) Core(TM) i7-8700K CPU @ 3.70GHz (skylake), Python 3.8.5
See https://gist.github.com/ed1de9ab3adf96a59f6ee68ba47690f6 for a full test report.

jhein32 · 2020-10-26T14:55:06Z

One of our users reported a floating point exception in MKL with v2020 u1. The issue is still present in v2020 u3. She was advised by intel that this is fixed in v2020 u4. We did a test install of that version and v2020 u4 resolved her issues.

I am preparing a PR to move MKL to 2020 u4

MKL component released in Oct 2020

move imkl to v2020.4.304 + bump version to intel/2020.10

boegel · 2020-10-27T20:47:14Z

Test report by @boegel
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in total)
node3163.skitty.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz, Python 3.6.8
See https://gist.github.com/e9a89aec64220f4c4fe4d241e2df1b1e for a full test report.

boegel · 2020-10-27T21:03:46Z

The problem that is causing the failing test on generoso can be fixed by defining $UCX_TLS, as follows:

export UCX_TLS=rc,ud,sm,self

See also the discussion in #10899 .

@bartoldeman Any idea if it's safe to always set this, or should I implement a hook on generoso to only inject that variable for specific UCX versions?

akesandgren · 2020-10-27T21:05:49Z

That's something that will be highly site specific, so please don't always set it.

bartoldeman · 2020-10-27T21:42:08Z

It's safe but not good for performance

jhein32 · 2020-10-27T21:58:23Z

When does it fail? We had issues with multi node running of MPI codes. Everything build fine.

I think UCX_TLS should be set in the Intel mpi module, since e.g. OpenMPI seems to handle this fine, when ÙCX-TLS`is unset. So setting this in the UCX modules for all MPI libs seems not the done thing to me.

akesandgren · 2020-10-28T05:39:15Z

It should not be set at all by upstream, as I said before, this is a site config and must remain so and, if at all, it should be handled in the site hooks.

boegel · 2020-10-28T07:11:45Z

@akesandgren Don't worry, I wasn't going to set it for everyone, just implement a hook on generoso to inject it into the impi module there...

generoso is a VM cluster that is only used for testing EasyBuild, MPI performance doesn't matter much there.

@jhein32 The problem occurs during the sanity check for impi, which is an mpirun of a trivial MPI hello world C program (the test/test.c that is included in the impi installation).

akesandgren · 2020-10-28T07:34:07Z

@boegel I know, it was more a comment to @jhein32

jhein32 · 2020-10-28T09:25:12Z

@boegel I know, it was more a comment to @jhein32

I understood, that disabling dc in general is a bad idea. We are singing from the same hymnsheet.

But for some of these things there are hints in easyconfigs what one might want do if things don't work. I suggested to put this into impi instead of the ucx module. But I realise that others here think differently.

jhein32 · 2020-10-28T09:30:03Z

@jhein32 The problem occurs during the sanity check for impi, which is an mpirun of a trivial MPI hello world C program (the test/test.c that is included in the impi installation).

@boegel Hmm, I never had that fail when we had issues with intel/2020a. I assume that test was in there already. On our system this tests inside the build node and will not use our stone-age IB. So if the non-existence of dc makes it trip, I am wondering whether it is going out of the (virtual) node. Does your "virtual cluster" have something resembling multiple nodes?

lexming · 2020-10-29T18:40:42Z

Test report by @lexming
SUCCESS
Build succeeded for 8 out of 8 (6 easyconfigs in total)
node101.hydra.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz, Python 2.7.5
See https://gist.github.com/f8468568e84b6b65cedd55a819ce6113 for a full test report.

boegel · 2020-11-06T21:42:05Z

I tested OpenFOAM 8 on top of intel/2020b after updating impi to 2019 update 9, works fine.

CP2K 7.1 test installation is still under way...

boegel · 2020-11-07T09:53:55Z

CP2K/7.1-intel-2020b worked fine on Intel Cascade Lake (CentOS 7), with good results for the regression test (correct: 3253 / 3270; new: 8; wrong: 8; failed: 1)

boegel · 2020-11-08T13:43:31Z

Also got WRF 3.9.1.1 to install with intel/2020b.

boegel · 2020-11-08T13:44:57Z

Test report by @boegel
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in total)
node3404.kirlia.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz (cascadelake), Python 2.7.5
See https://gist.github.com/2d2567a33db9ad87f83f2b28970e0cde for a full test report.

boegel · 2020-11-08T13:45:41Z

Test report by @boegel
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in total)
node3108.skitty.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz, Python 3.6.8
See https://gist.github.com/4a9101ca528d0069c3d70c50a0ec5a58 for a full test report.

Micket · 2020-11-08T23:13:00Z

Test report by @Micket
SUCCESS
Build succeeded for 6 out of 6 (6 easyconfigs in total)
vera-c1 - Linux centos linux 7.8.2003, x86_64, Intel Xeon Processor (Skylake), Python 2.7.5
See https://gist.github.com/c93a2ef57c8e521481e966529e56bb31 for a full test report.

Micket

lgtm

Micket · 2020-11-09T09:43:30Z

Going in, thanks @boegel!

{toolchain} intel/2020.09

aed55bf

boegel added the update label Sep 23, 2020

use https download URLs latest iccifort and imkl

3586076

boegel added this to the next release (4.3.1) milestone Sep 23, 2020

easybuilders deleted a comment from boegelbot Sep 23, 2020

boegel added the 2020b issues & PRs related to 2020b label Sep 24, 2020

lexming reviewed Sep 25, 2020

View reviewed changes

This was referenced Sep 25, 2020

IMPI v2019.6: MLX provider in libfabric not working #10213

Open

Problem with multinode running with iimpi-2020a #10899

Closed

remove suggestion to define $I_MPI_HYDRA_TOPOLIB as 'ipl' to work aro…

f1016bf

…und impi bug, no longer relevant for impi 2019 update 8 Co-authored-by: Alex Domingo <alex.domingo.toro@vub.be>

move imkl to v 2020.4.304

b65f984

jhein32 and others added 2 commits October 26, 2020 16:25

Rename toolchain from 2020.09 to 2020.10

624961e

MKL component released in Oct 2020

Merge pull request #77 from jhein32/imkl_2020_u4_update

dc6ac78

move imkl to v2020.4.304 + bump version to intel/2020.10

boegel mentioned this pull request Oct 26, 2020

move imkl to v2020.4.304 + bump version to intel/2020.10 boegel/easybuild-easyconfigs#77

Merged

boegel changed the title ~~{toolchain} intel/2020.09~~ {toolchain} intel/2020.10 (candidate for intel/2020b) [WIP] Oct 26, 2020

update intel/2020.10 to Intel compilers 2020 update 4

90a131c

boegel added 2 commits November 6, 2020 09:01

update impi to 2019 update 9 in intel/2020.10 toolchain

6e1f523

rename intel/2020.10 to intel/2020b

a3b4f88

boegel changed the title ~~{toolchain} intel/2020.10 (candidate for intel/2020b) [WIP]~~ {toolchain} intel/2020b Nov 6, 2020

boegel mentioned this pull request Nov 8, 2020

{lang}[foss/2020b,intel/2020b] SciPy-bundle v2020.11 w/ Python 3.8.6 #11629

Merged

This comment has been minimized.

Sign in to view

Micket approved these changes Nov 9, 2020

View reviewed changes

Micket merged commit 3ee8e5a into easybuilders:develop Nov 9, 2020

boegel deleted the 20200923141429_new_pr_HPL23 branch November 9, 2020 10:08

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

{toolchain} intel/2020b #11337

{toolchain} intel/2020b #11337

boegel commented Sep 23, 2020 •

edited

Loading

boegel commented Sep 23, 2020

boegelbot commented Sep 23, 2020

boegelbot commented Sep 23, 2020 •

edited by boegel

Loading

boegel commented Sep 23, 2020

lexming commented Sep 24, 2020

lexming commented Sep 24, 2020

boegel commented Oct 5, 2020

boegelbot commented Oct 5, 2020

boegelbot commented Oct 5, 2020 •

edited by boegel

Loading

jhein32 commented Oct 9, 2020

boegel commented Oct 9, 2020 •

edited

Loading

jhein32 commented Oct 12, 2020

zao commented Oct 16, 2020

jhein32 commented Oct 26, 2020

boegel commented Oct 27, 2020

boegel commented Oct 27, 2020

akesandgren commented Oct 27, 2020

bartoldeman commented Oct 27, 2020

jhein32 commented Oct 27, 2020

akesandgren commented Oct 28, 2020

boegel commented Oct 28, 2020

akesandgren commented Oct 28, 2020

jhein32 commented Oct 28, 2020

jhein32 commented Oct 28, 2020

lexming commented Oct 29, 2020

boegel commented Nov 6, 2020

boegel commented Nov 7, 2020

boegel commented Nov 8, 2020

boegel commented Nov 8, 2020

boegel commented Nov 8, 2020

This comment has been minimized.

Micket commented Nov 8, 2020

Micket left a comment

Micket commented Nov 9, 2020

{toolchain} intel/2020b #11337

{toolchain} intel/2020b #11337

Conversation

boegel commented Sep 23, 2020 • edited Loading

boegel commented Sep 23, 2020

boegelbot commented Sep 23, 2020

boegelbot commented Sep 23, 2020 • edited by boegel Loading

boegel commented Sep 23, 2020

lexming commented Sep 24, 2020

lexming commented Sep 24, 2020

boegel commented Oct 5, 2020

boegelbot commented Oct 5, 2020

boegelbot commented Oct 5, 2020 • edited by boegel Loading

jhein32 commented Oct 9, 2020

boegel commented Oct 9, 2020 • edited Loading

jhein32 commented Oct 12, 2020

zao commented Oct 16, 2020

jhein32 commented Oct 26, 2020

boegel commented Oct 27, 2020

boegel commented Oct 27, 2020

akesandgren commented Oct 27, 2020

bartoldeman commented Oct 27, 2020

jhein32 commented Oct 27, 2020

akesandgren commented Oct 28, 2020

boegel commented Oct 28, 2020

akesandgren commented Oct 28, 2020

jhein32 commented Oct 28, 2020

jhein32 commented Oct 28, 2020

lexming commented Oct 29, 2020

boegel commented Nov 6, 2020

boegel commented Nov 7, 2020

boegel commented Nov 8, 2020

boegel commented Nov 8, 2020

boegel commented Nov 8, 2020

This comment has been minimized.

Micket commented Nov 8, 2020

Micket left a comment

Choose a reason for hiding this comment

Micket commented Nov 9, 2020

boegel commented Sep 23, 2020 •

edited

Loading

boegelbot commented Sep 23, 2020 •

edited by boegel

Loading

boegelbot commented Oct 5, 2020 •

edited by boegel

Loading

boegel commented Oct 9, 2020 •

edited

Loading