(WIP) added CUDA 11.0.2 and related recipes #10935

mboisson · 2020-07-07T20:35:18Z

No description provided.

easybuild/easyconfigs/c/CUDAcore/CUDAcore-11.0.2.eb

easybuild/easyconfigs/c/CUDA/CUDA-11.0-iccifort-2020.1.217.eb

mboisson · 2020-07-07T21:28:17Z

I'm not sure what this failure means

Traceback (most recent call last):
  File "/home/runner/work/easybuild-easyconfigs/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 773, in test_changed_files_pull_request
    self.check_sanity_check_paths(changed_ecs)
  File "/home/runner/work/easybuild-easyconfigs/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 650, in check_sanity_check_paths
    self.assertFalse(failing_checks, '\n'.join(failing_checks))
AssertionError: ['No custom sanity_check_paths found in CUDA-11.0-GCC-9.3.0.eb', 'No custom sanity_check_paths found in CUDA-11.0-iccifort-2020.1.217.eb'] is not false : No custom sanity_check_paths found in CUDA-11.0-GCC-9.3.0.eb
No custom sanity_check_paths found in CUDA-11.0-iccifort-2020.1.217.eb

there are no sanity_check_paths in any of the other CUDA recipes ?

mboisson · 2020-07-07T21:31:58Z

Ah, it's because it's using a Bundle EasyBlock, which is not in here :

easybuild-easyconfigs/test/easyconfigs/easyconfigs.py

Line 632 in 48c6786

    
           whitelist = ['BuildEnv', 'CrayToolchain', 'GoPackage', 'ModuleRC', 'PythonBundle', 'PythonPackage',

Why are PythonBundle excluded, but not Bundle ?

Micket · 2020-07-07T21:37:04Z

Test report by @Micket
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in this PR)
hebbe-c1 - Linux centos linux 7.8.2003, x86_64, Intel Core Processor (Haswell, no TSX), Python 2.7.5
See https://gist.github.com/4c6265827ff87a3f08452518f6b763c0 for a full test report.

(this was a testbuilt using the fixed cuda.py block)

Micket · 2020-07-07T21:42:31Z

Why are PythonBundle excluded, but not Bundle ?

PythonPackage and PythonBundle are excluded because they contain default implementations for sanity checks. It checks for the site-packages directory
Bundles typically always contain stuff though, this is really a.. uh. well I don't know really. The only cases that are module-only are actually ModuleRC (which wouldn't work) and Toolchain as far as I know. (oh, and BuildEnv, but that's not right either)

mboisson · 2020-07-07T22:03:49Z

Why are PythonBundle excluded, but not Bundle ?

PythonPackage and PythonBundle are excluded because they contain default implementations for sanity checks. It checks for the site-packages directory
Bundles typically always contain stuff though, this is really a.. uh. well I don't know really. The only cases that are module-only are actually ModuleRC (which wouldn't work) and Toolchain as far as I know. (oh, and BuildEnv, but that's not right either)

I can try to create an empty sanity_check_paths ?

mboisson · 2020-07-07T22:04:35Z

Ah, no, I need to add to

easybuild-easyconfigs/test/easyconfigs/easyconfigs.py

Line 635 in 48c6786

bundles_whitelist = ['Autotools', 'GCC']

mboisson · 2020-07-07T22:11:28Z

Ah, no, I need to add to

easybuild-easyconfigs/test/easyconfigs/easyconfigs.py

Line 635 in 48c6786

bundles_whitelist = ['Autotools', 'GCC']

That is done in
#10936

Micket · 2020-07-07T22:41:33Z

(this is still the most reliable way to re-test a PR)

Micket · 2020-07-07T22:41:44Z

Re-running tests

mboisson · 2020-07-08T13:43:27Z

I'm not sure why the tests fail with

Traceback (most recent call last):
  File "/home/runner/work/easybuild-easyconfigs/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 147, in test_conflicts
    self.process_all_easyconfigs()
  File "/home/runner/work/easybuild-easyconfigs/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 120, in process_all_easyconfigs
    EasyConfigTest.ordered_specs = resolve_dependencies(EasyConfigTest.parsed_easyconfigs, modules_tool(), retain_all_deps=True)
  File "/opt/hostedtoolcache/Python/3.5.9/x64/lib/python3.5/site-packages/easybuild/tools/robot.py", line 460, in resolve_dependencies
    raise_error_missing_deps(totally_missing, extra_msg="no easyconfig file or existing module found")
  File "/opt/hostedtoolcache/Python/3.5.9/x64/lib/python3.5/site-packages/easybuild/tools/robot.py", line 323, in raise_error_missing_deps
    raise EasyBuildError(error_msg)
easybuild.tools.build_log.EasyBuildError: 'Missing dependencies: CUDAcore/11.0.2-GCC-9.3.0, CUDAcore/11.0.2-iccifort-2020.1.217 (no easyconfig file or existing module found)'

The recipes do not call for CUDAcore/11.0.2-GCC-9.3.0. Shouldn't EasyBuild automatically resolve the dependencies are at the SYSTEM level ?

mboisson · 2020-07-08T17:38:08Z

Looks like the build is simply too long ?

bartoldeman · 2020-07-08T19:16:07Z

Travis has issues from time to time. I'll try again.

bartoldeman · 2020-07-08T23:14:03Z

UCX should be built with CUDA support too, Open MPI in here is using the non-CUDA enabled UCX. But that would need the gcccorecuda toolchain as explained in slack.

mboisson · 2020-07-09T00:06:19Z

Ah.... so we need to do a PR for easyblocks ? Is it just the new toolchain, or something else ?

…

On Jul 8, 2020 at 7:14 PM, <bartoldeman ***@***.***)> wrote: UCX should be built with CUDA support too, Open MPI in here is using the non-CUDA enabled UCX. But that would need the gcccorecuda toolchain as explained in slack. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub (#10935 (comment)), or unsubscribe (https://github.com/notifications/unsubscribe-auth/ABZKY2U7DLDD7FTK73NEQADR2T4URANCNFSM4OTR7GNQ).

bartoldeman · 2020-07-09T13:33:58Z

No new easyblock is needed, but a gcccorecuda toolchain needs to be added to the framework.

Framework PR filed here: easybuilders/easybuild-framework#3385

Used for UCX for example which uses GCCcore for the non-CUDA version.

This reverts commit 34c6bb2.

This reverts commit ee5ee6d.

boegelbot · 2020-07-09T17:18:16Z

Travis test report: 2/2 runs failed - see https://travis-ci.org/easybuilders/easybuild-easyconfigs/builds/706594322

Only showing partial log for 1st failed test suite run 23191.1;
full log at https://travis-ci.org/easybuilders/easybuild-easyconfigs/jobs/706594323

...
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/travis/build/easybuilders/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 492, in test_sanity_check_paths
    self.process_all_easyconfigs()
  File "/home/travis/build/easybuilders/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 120, in process_all_easyconfigs
    EasyConfigTest.ordered_specs = resolve_dependencies(EasyConfigTest.parsed_easyconfigs, modules_tool(), retain_all_deps=True)
  File "/home/travis/virtualenv/python2.7.15/lib/python2.7/site-packages/easybuild/tools/robot.py", line 436, in resolve_dependencies
    processed_ecs = process_easyconfig(path, validate=not retain_all_deps, hidden=hidden)
  File "/home/travis/virtualenv/python2.7.15/lib/python2.7/site-packages/easybuild/framework/easyconfig/easyconfig.py", line 1966, in process_easyconfig
    raise EasyBuildError("Failed to process easyconfig %s: %s", spec, err.msg)
EasyBuildError: 'Failed to process easyconfig /home/travis/build/easybuilders/easybuild-easyconfigs/easybuild/easyconfigs/u/UCX/UCX-1.8.0-gcccorecuda-2020a.eb: Toolchain gcccorecuda not found, available toolchains: giolfc,gpsmpi,intel,iomkl,dummy,gimpic,goblf,pompi,cgmpich,gomkl,ismkl,iimpi,system,gsmpi,xlmvapich2,cgmpolf,gmpich2,cgoolf,gqacml,CrayPGI,gmvapich2,iimkl,impich,iiqmpi,xlmpich2,foss,iompi,goolf,iomklc,GCCcore,gmpich,xlcxlf,golf,iompic,giolf,gcccuda,cgmvapich2,iimklc,xlompi,gompi,gomklc,CrayCCE,iccifort,intel-para,ipsmpi,gmpolf,GNU,iccifortcuda,ictce,gmkl,goalf,goolfc,gimpi,gmvolf,intelcuda,gmklc,CrayIntel,impmkl,iqacml,CrayGNU,golfc,gpsolf,cgmvolf,GCC,iimpic,ClangGCC,xlmpich,pomkl,gmacml,gimkl,cgompi,PGI,fosscuda,pmkl,gsolf,gompic'

======================================================================
FAIL: test_changed_files_pull_request (test.easyconfigs.easyconfigs.EasyConfigTest)
Specific checks only done for the (easyconfig) files that were changed in a pull request.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/travis/build/easybuilders/easybuild-easyconfigs/test/easyconfigs/easyconfigs.py", line 768, in test_changed_files_pull_request
    self.assertTrue(False, error_msg)
AssertionError: Failed to find parsed easyconfig for UCX-1.8.0-gcccorecuda-2020a.eb (and could not isolate it in easyconfigs archive either)

----------------------------------------------------------------------
Ran 10170 tests in 365.249s

FAILED (failures=1, errors=3)
ERROR: Not all tests were successful.
travis_time:end:295feb58:start=1594312504382178034,finish=1594312871815642595,duration=367433464561,event=script
�[0K�[31;1mThe command "python -O -m test.easyconfigs.suite" exited with 2.�[0m
travis_time:start:132030ec
�[0K$ unset PYTHONPATH
travis_time:end:132030ec:start=1594312871819568119,finish=1594312871821850615,duration=2282496,event=script
�[0K�[32;1mThe command "unset PYTHONPATH" exited with 0.�[0m
travis_time:start:05f81a10
�[0K$ cd $HOME; pip install $TRAVIS_BUILD_DIR
�[33mDEPRECATION: Python 2.7 will reach the end of its life on January 1st, 2020. Please upgrade your Python as Python 2.7 won't be maintained after that date. A future version of pip will drop support for Python 2.7.�[0m
Processing ./build/easybuilders/easybuild-easyconfigs
Building wheels for collected packages: easybuild-easyconfigs
  Building wheel for easybuild-easyconfigs (setup.py) ... �[?25l-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �|� �/� �-� �\� �done
�[?25h  Stored in directory: /home/travis/.cache/pip/wheels/60/83/5a/68a83b743a3b5a96bdb88de299d7630c9f272b0dec140d3e5d
Successfully built easybuild-easyconfigs
Installing collected packages: easybuild-easyconfigs
Successfully installed easybuild-easyconfigs-4.2.3.dev0
travis_time:end:05f81a10:start=1594312871825472870,finish=1594312882847023922,duration=11021551052,event=script
�[0K�[32;1mThe command "cd $HOME; pip install $TRAVIS_BUILD_DIR" exited with 0.�[0m
travis_time:start:074c0de0
�[0K$ export EB_PYTHON=python
travis_time:end:074c0de0:start=1594312882850906920,finish=1594312882853244136,duration=2337216,event=script
�[0K�[32;1mThe command "export EB_PYTHON=python" exited with 0.�[0m
travis_time:start:0911d460
�[0K$ eb --show-config | tee eb_show_config.out
#
# Current EasyBuild configuration
# (C: command line argument, D: default value, E: environment variable, F: configuration file)
#
buildpath      (D) = /home/travis/.local/easybuild/build
containerpath  (D) = /home/travis/.local/easybuild/containers
installpath    (D) = /home/travis/.local/easybuild
repositorypath (D) = /home/travis/.local/easybuild/ebfiles_repo
robot-paths    (D) = /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs
sourcepath     (D) = /home/travis/.local/easybuild/sources
travis_time:end:0911d460:start=1594312882856603120,finish=1594312883266044269,duration=409441149,event=script
�[0K�[32;1mThe command "eb --show-config | tee eb_show_config.out" exited with 0.�[0m
travis_time:start:17f8f716
�[0K$ grep "^robot-paths .*/easybuild/easyconfigs" eb_show_config.out
robot-paths    (D) = /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs
travis_time:end:17f8f716:start=1594312883270205959,finish=1594312883274280797,duration=4074838,event=script
�[0K�[32;1mThe command "grep "^robot-paths .*/easybuild/easyconfigs" eb_show_config.out" exited with 0.�[0m
travis_time:start:1b46f094
�[0K$ eb --search 'TensorFlow-1.14.*.eb' | tee eb_search_TF.out
 * /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs/t/TensorFlow/TensorFlow-1.14.0-foss-2019a-Python-3.7.2.eb
 * /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs/t/TensorFlow/TensorFlow-1.14.0-fosscuda-2019a-Python-3.7.2.eb
travis_time:end:1b46f094:start=1594312883277714704,finish=1594312883862777650,duration=585062946,event=script
�[0K�[32;1mThe command "eb --search 'TensorFlow-1.14.*.eb' | tee eb_search_TF.out" exited with 0.�[0m
travis_time:start:08bdba4c
�[0K$ grep '/TensorFlow-1.14.0-foss-2019a-Python-3.7.2.eb$' eb_search_TF.out
 * /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs/t/TensorFlow/TensorFlow-1.14.0-foss-2019a-Python-3.7.2.eb
travis_time:end:08bdba4c:start=1594312883867063467,finish=1594312883871076954,duration=4013487,event=script
�[0K�[32;1mThe command "grep '/TensorFlow-1.14.0-foss-2019a-Python-3.7.2.eb$' eb_search_TF.out" exited with 0.�[0m
travis_time:start:005a60f4
�[0K$ eb --search '^foss-2018b.eb' | tee eb_search_foss.out
 * /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs/f/foss/foss-2018b.eb
travis_time:end:005a60f4:start=1594312883874741798,finish=1594312884452892082,duration=578150284,event=script
�[0K�[32;1mThe command "eb --search '^foss-2018b.eb' | tee eb_search_foss.out" exited with 0.�[0m
travis_time:start:17cba45f
�[0K$ grep '/foss-2018b.eb$' eb_search_foss.out
 * /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs/f/foss/foss-2018b.eb
travis_time:end:17cba45f:start=1594312884457223077,finish=1594312884460938867,duration=3715790,event=script
�[0K�[32;1mThe command "grep '/foss-2018b.eb$' eb_search_foss.out" exited with 0.�[0m
travis_time:start:00dea239
�[0K$ eb --prefix /tmp/$USER M4-1.4.18.eb
== temporary log file in case of crash /tmp/eb-_YdiHf/easybuild-mM1Gw8.log
== processing EasyBuild easyconfig /home/travis/virtualenv/python2.7.15/easybuild/easyconfigs/m/M4/M4-1.4.18.eb
== building and installing M4/1.4.18...
== fetching files...
== creating build dir, resetting environment...
== unpacking...
== patching...
== preparing...
== configuring...
== building...
== testing...
== installing...
== taking care of extensions...
== restore after iterating...
== postprocessing...
== sanity checking...
== cleaning up...
== creating module...
== permissions...
== packaging...
== COMPLETED: Installation ended successfully (took 25 sec)
== Results of the build can be found in the log file(s) /tmp/travis/software/M4/1.4.18/easybuild/easybuild-M4-1.4.18-20200709.164149.log
== Build succeeded for 1 out of 1
== Temporary log file(s) /tmp/eb-_YdiHf/easybuild-mM1Gw8.log* have been removed.
== Temporary directory /tmp/eb-_YdiHf has been removed.
travis_time:end:00dea239:start=1594312884464383727,finish=1594312909920512406,duration=25456128679,event=script
�[0K�[32;1mThe command "eb --prefix /tmp/$USER M4-1.4.18.eb" exited with 0.�[0m


Done. Your build exited with 1.

23191.2 - failed => https://travis-ci.org/easybuilders/easybuild-easyconfigs/jobs/706594324

*bleep, bloop, I'm just a bot (boegelbot v20180813.01)*Please talk to my owner @boegel if you notice you me acting stupid),or submit a pull request to https://github.com/boegel/boegelbot fix the problem.

boegel · 2020-08-19T15:36:06Z

easybuild/easyconfigs/o/OpenMPI/OpenMPI-4.0.3-iccifortcuda-2020a.eb

+dependencies = [
+    ('zlib', '1.2.11'),
+    ('hwloc', '2.2.0'),
+    ('UCX', '1.8.0', '', ('gcccorecuda', '2020a')),


There should be a comment here to explain why we do this.

Like easybuilders#10935 this introduces a new CUDAcore easyconfig to share CUDA between use of GCC and Intel compilers, but unlike easybuilders#10935 this uses a versionsuffix for UCX+CUDA so it does not need any framework changes or MODULEPATH adjustments.

mboisson · 2022-10-13T17:06:14Z

This is no longer relevant I guess.

added CUDA 11.0.2 and related recipes

7c3581d

Micket requested changes Jul 7, 2020

View reviewed changes

easybuild/easyconfigs/c/CUDAcore/CUDAcore-11.0.2.eb Show resolved Hide resolved

easybuild/easyconfigs/c/CUDA/CUDA-11.0-iccifort-2020.1.217.eb Outdated Show resolved Hide resolved

Micket added change update and removed change labels Jul 7, 2020

Micket added this to the release after 4.2.2 (4.2.3?) milestone Jul 7, 2020

added support for ppc64le

a0acd5b

Micket closed this Jul 7, 2020

Micket reopened this Jul 7, 2020

This comment has been minimized.

Sign in to view

mboisson added 3 commits July 8, 2020 13:43

fixed coding style

e3241db

fixing coding style in CUDAcore 11.0.2

bcd017a

allowed CUDA to be used from the SYSTEM toolchain

a980e3c

mboisson closed this Jul 8, 2020

mboisson reopened this Jul 8, 2020

bartoldeman closed this Jul 8, 2020

bartoldeman reopened this Jul 8, 2020

restored revision number for CUDA

343e208

mboisson added 3 commits July 8, 2020 20:39

added OpenMPI-4.0.3-gcccuda-2020a.eb

be1f650

added OpenMPI-4.0.3-iccifortcuda-2020a.eb

80ea4f2

added toolchain gompic-2020a

80885f6

bartoldeman and others added 6 commits July 9, 2020 13:53

Add gcccorecuda for 2020a toolchains.

69b8500

Used for UCX for example which uses GCCcore for the non-CUDA version.

UCX: add 1.8.0 with CUDA 11.0 support

3221c5b

libfabric: compile with CUDA 11

ee5ee6d

libfabric: need explicit CUDA_ROOT, otherwise it will use "yes"

34c6bb2

Revert "libfabric: need explicit CUDA_ROOT, otherwise it will use "yes""

b61846b

This reverts commit 34c6bb2.

Revert "libfabric: compile with CUDA 11"

2c77aeb

This reverts commit ee5ee6d.

This comment has been minimized.

Sign in to view

resolving the ambiguity between cuda and non-cuda UCX

cf72904

mboisson mentioned this pull request Jul 9, 2020

(WIP) {g,i}omklc toolchains and relatives, version 2020a #10954

Closed

boegel requested changes Aug 19, 2020

View reviewed changes

boegel modified the milestones: next release (4.3.0), release after 4.3.0 (4.3.1?) Sep 9, 2020

bartoldeman mentioned this pull request Sep 16, 2020

Added CUDA 11.0.2 and related recipes, incl. gompic/2020a and iccifortcuda/2020a #11295

Merged

boegel modified the milestones: 4.3.1 (next release), 4.x Oct 26, 2020

mboisson closed this Oct 13, 2022

mboisson deleted the cuda-11.0.2 branch October 13, 2022 17:07

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

(WIP) added CUDA 11.0.2 and related recipes #10935

(WIP) added CUDA 11.0.2 and related recipes #10935

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020 •

edited

Loading

Micket commented Jul 7, 2020 •

edited

Loading

Micket commented Jul 7, 2020 •

edited

Loading

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020

Micket commented Jul 7, 2020

Micket commented Jul 7, 2020

This comment has been minimized.

mboisson commented Jul 8, 2020

mboisson commented Jul 8, 2020

bartoldeman commented Jul 8, 2020

bartoldeman commented Jul 8, 2020

mboisson commented Jul 9, 2020 via email

bartoldeman commented Jul 9, 2020

This comment has been minimized.

boegelbot commented Jul 9, 2020

boegel Aug 19, 2020

mboisson commented Oct 13, 2022

(WIP) added CUDA 11.0.2 and related recipes #10935

(WIP) added CUDA 11.0.2 and related recipes #10935

Conversation

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020 • edited Loading

Micket commented Jul 7, 2020 • edited Loading

Micket commented Jul 7, 2020 • edited Loading

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020

mboisson commented Jul 7, 2020

Micket commented Jul 7, 2020

Micket commented Jul 7, 2020

This comment has been minimized.

mboisson commented Jul 8, 2020

mboisson commented Jul 8, 2020

bartoldeman commented Jul 8, 2020

bartoldeman commented Jul 8, 2020

mboisson commented Jul 9, 2020 via email

bartoldeman commented Jul 9, 2020

This comment has been minimized.

boegelbot commented Jul 9, 2020

boegel Aug 19, 2020

Choose a reason for hiding this comment

mboisson commented Oct 13, 2022

mboisson commented Jul 7, 2020 •

edited

Loading

Micket commented Jul 7, 2020 •

edited

Loading

Micket commented Jul 7, 2020 •

edited

Loading