Hi Kenneth and Jan, Thanks for the rapid response. I was able to do the installation with the PR without any problems.
Cheers, Loris Kenneth Hoste <[email protected]> writes: > Hi all, > > Just to add to this: we'll be pushing out an additional quick > EasyBuild v4.9.4 bugfix release because of this, since it's likely > that many people will be hitting the problem reported by Loris. > > Despite rigorous pre-release regression testing, in which we literally > run thousands of software installations with what will be the next > EasyBuild release, this one slipped between the cracks... > > In the main test environment that I was using for the regression test, > the cuda-compute-capabilities EasyBuild configuration option was set, > and hence the problem reported below was not observed at all. > > Same goes for the additional test environments in which the pull > request that introduces this problem was extensive tested (see > https://github.com/easybuilders/easybuild-easyblocks/pull/3396). > > As the pull request that fixes this problem > (https://github.com/easybuilders/easybuild-easyblocks/pull/3449) > shows, the bug was also quite easy to overlook when reviewing the > changes. > > I expect to release EasyBuild v4.9.4 in the coming days that will > include the fix for this problem. > > Until then, as Jan already mentioned, you can use > "--include-easyblocks-from-pr 3449" as a workaround for this issue. > > regards, > > Kenneth > > On 18/09/2024 14:24, Jan André Reuter wrote: >> Heya, >> unfortunately, a change to the GCCcore EasyBlock, which was meant to >> improve default NVPTX architecture for GCC 13.1 and newer, has >> introduced this issue. In EasyBuild installations without a CUDA >> compute capability set, the code fails to check for an empty list, >> causing this error to show up. >> For now, "|eb --include-easyblocks-from-pr 3449" can be used as a >> |workaround|. This PR fixes the bug in the EasyBlock.| >> |Best regards, >> Jan >> ||| >> On 18.09.24 14:07, Loris Bennett wrote: >>> Hi, >>> >>> With EasyBuild 4.9.3 and >>> >>> eb Rust-1.79.0-GCCcore-13.3.0.eb --robot --parallel=8 >>> >>> I get >>> >>> == building and installing GCCcore/13.3.0... >>> ... >>> ERROR: Traceback (most recent call last): >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/main.py", >>> line 137, in build_and_install_software >>> (ec_res['success'], app_log, err) = build_and_install_one(ec, >>> init_env) >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/framework/easyblock.py", >>> line 4276, in build_and_install_one >>> result = app.run_all_steps(run_test_cases=run_test_cases) >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/easyblocks/g/gcc.py", >>> line 1081, in run_all_steps >>> return super(EB_GCC, self).run_all_steps(*args, **kwargs) >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/framework/easyblock.py", >>> line 4155, in run_all_steps >>> self.run_step(step_name, step_methods) >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/framework/easyblock.py", >>> line 3990, in run_step >>> step_method(self)() >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/easyblocks/g/gcc.py", >>> line 612, in configure_step >>> cuda_cc = self.map_nvptx_capability() >>> File >>> "/trinity/shared/easybuild/software/EasyBuild/4.9.3/lib/python3.6/site-packages/easybuild/easyblocks/g/gcc.py", >>> line 431, in map_nvptx_capability >>> return sorted_gcc_cc[0] >>> IndexError: list index out of range >>> >>> Looks like the list >>> >>> sorted_gcc_cc >>> >>> is empty. Does anyone know why this might be? >>> >>> Cheers, >>> >>> Loris >>> >> -- --- >> Jan André Reuter >> Jülich Supercomputing Centre (JSC) >> ATML Parallel Performance >> Phone: +49-2461-61-8871 >> E-Mail:[email protected] >> Internet:http://www.fz-juelich.de >> --------------------------------------------------------------------------------------------- >> --------------------------------------------------------------------------------------------- >> Forschungszentrum Jülich GmbH >> 52425 Jülich >> Sitz der Gesellschaft: Jülich >> Eingetragen im Handelsregister des Amtsgerichts Düren Nr. HR B 3498 >> Vorsitzender des Aufsichtsrats: MinDir Stefan Müller >> Geschäftsführung: Prof. Dr. Astrid Lambrecht (Vorsitzende), >> Karsten Beneke (stellv. Vorsitzender), Dr. Ir. Pieter Jansens >> --------------------------------------------------------------------------------------------- >> --------------------------------------------------------------------------------------------- >> > -- Dr. Loris Bennett (Herr/Mr) FUB-IT, Freie Universität Berlin
