{math}[foss/2019b,fosscuda/2019b] Keras v2.3.1, TensorFlow v2.1.0, libgpuarray v0.7.6, Theano v1.0.4, R-keras v2.2.5.0 w/ Python 3.7.4

[edmondac]

(created using eb --new-pr)

This depends on #9519 being merged first

[edmondac]

That patch is to make Bazel compile on my Ubuntu 19.10 machine. It comes from https://github.com/clearlinux-pkgs/bazel/blob/adefd9046582cb52f39579033132e6265ef6ddb0/rename-gettid-functions.patch

[terjekv]

Test report by @terjekv
SUCCESS
Build succeeded for 10 out of 10 (8 easyconfigs in this PR)
ninhursaga.uio.no - Linux RHEL 8.1, Intel(R) Xeon(R) Silver 4216 CPU @ 2.10GHz, Python 3.6.8
See https://gist.github.com/2faeb8881f5a76da5642a0d65ab7aea3 for a full test report.

[edmondac]

Test report by @bear-rsg
SUCCESS
Build succeeded for 9 out of 9 (8 easyconfigs in this PR)
bear-pg0212u15a.bear.cluster - Linux centos linux 7.7.1908, Intel(R) Xeon(R) CPU E5-2640 v4 @ 2.40GHz, Python 2.7.5
See https://gist.github.com/b6f8e6181b7d47323825530f3d34e6f0 for a full test report.

[boegel]

Test report by @boegel
FAILED
Build succeeded for 9 out of 11 (8 easyconfigs in this PR)
node3170.skitty.os - Linux centos linux 7.7.1908, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz, Python 3.6.8
See https://gist.github.com/e25f0dc36f832b6b6af19edd74efb278 for a full test report.

[terjekv]

Test report by @boegel
FAILED
Build succeeded for 9 out of 11 (8 easyconfigs in this PR)
node3170.skitty.os - Linux centos linux 7.7.1908, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz, Python 3.6.8
See https://gist.github.com/e25f0dc36f832b6b6af19edd74efb278 for a full test report.

Checksum failure for R packages? Well I never... Shocking!

[Flamefire]

Can we have an --add-alternate-checksum option to EB? Until then: @edmondac Can you just add that checksum and double-check with deleted archives?

[boegel]

@edmondac See fixed checksum in #9879

[edmondac]

@boegel that should do it, I think

[Flamefire]

Running test on our system now.
Edit: Restarted due to FS failure

[edmondac]

Test report by @bear-rsg
SUCCESS
Build succeeded for 11 out of 11 (8 easyconfigs in this PR)
bear-pg0212u15a.bear.cluster - Linux centos linux 7.7.1908, Intel(R) Xeon(R) CPU E5-2640 v4 @ 2.40GHz, Python 2.7.5
See https://gist.github.com/10654d29e0daa0fc88eabfaaaad31e04 for a full test report.

[Flamefire]

Test report by @Flamefire
SUCCESS
Build succeeded for 8 out of 8 (8 easyconfigs in this PR)
taurusi6325.taurus.hrsk.tu-dresden.de - Linux RHEL 7.7, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz, Python 2.7.5
See https://gist.github.com/b75473ee2418bec25dc0812ecc6f655d for a full test report.

[edmondac]

OK @boegel we think we're happy with this :-)

[terjekv]

Test report by @Flamefire
FAILED
Build succeeded for 7 out of 10 (8 easyconfigs in this PR)
taurusml30 - Linux RHEL 7.6, 8335-GTX, Python 2.7.5
See https://gist.github.com/fa43829e112ba485821ce5bda243d52e for a full test report.

Hm, an issue with Java/1.8 on power9?

== 2020-02-26 18:59:06,386 build_log.py:169 ERROR EasyBuild crashed with an error (at easybuild/base/exceptions.py:124 in __init__): Module command 'module load Bazel/0.29.1-GCCcore-8.3.0' failed with exit code 1; stderr: Lmod hat den folgenden Fehler erkannt: Modul konnte nicht geladen werden: ""
     /sw/modules/ml/devel/Bazel/0.29.1-GCCcore-8.3.0.lua: /usr/share/lmod/lmod/libexec/MasterControl.lua:888: attempt to compare number with nil

Executing this command requires loading "Java/1.8" which failed while processing the following module(s):

    Module fullname             Module Filename
    ---------------             ---------------
    Bazel/0.29.1-GCCcore-8.3.0  /sw/modules/ml/devel/Bazel/0.29.1-GCCcore-8.3.0.lua
While processing the following module(s):
    Module fullname             Module Filename
    ---------------             ---------------
    Bazel/0.29.1-GCCcore-8.3.0  /sw/modules/ml/devel/Bazel/0.29.1-GCCcore-8.3.0.lua

; stdout: 
false

[Flamefire]

This seems to be an LMod 7(.8.9) issue where it can't find the proper module alias. Retrying with latest LMod

[terjekv]

Oh, yeah, right. That was brought up on Slack I think. Hopefully it's an easy LMod upgrade away from working.

[Flamefire]

Turns out to be a real LMod bug: TACC/Lmod#436

I have a work-around and it's now starting to build TensorFlow. Going to take a while...

[Flamefire]

Test report by @Flamefire
SUCCESS
Build succeeded for 8 out of 8 (8 easyconfigs in this PR)
taurusml13 - Linux RHEL 7.6, 8335-GTX, Python 2.7.5
See https://gist.github.com/6eb4e41b1b52aa648189fd405ca27a15 for a full test report.

[Flamefire]

Ok with the patches for R and the like (see my PRs) this works on Power9. So 👍 from me!

[terjekv]

Test report by @terjekv
SUCCESS
Build succeeded for 12 out of 12 (8 easyconfigs in this PR)
ninhursaga.uio.no - Linux RHEL 8.1, Intel(R) Xeon(R) Silver 4216 CPU @ 2.10GHz, Python 3.6.8
See https://gist.github.com/1c662b21205fa30e3eabb5b262012246 for a full test report.

[boegel]

Test report by @boegel
SUCCESS
Build succeeded for 10 out of 10 (8 easyconfigs in this PR)
node3306.joltik.os - Linux centos linux 7.7.1908, Intel(R) Xeon(R) Gold 6242 CPU @ 2.80GHz, Python 3.6.8
See https://gist.github.com/bba1c53c820776a56561ae232c20040d for a full test report.

[boegel]

lgtm

[boegel]

Going in, thanks @edmondac!