Add configuration for AMD to EESSI-extend-easybuild.eb by zerefwayne · Pull Request #206 · EESSI/software-layer-scripts

zerefwayne · 2026-04-17T11:40:05Z

No description provided.

ocaisa

LGTM but we will need to add some testing for this in https://github.com/EESSI/software-layer-scripts/blob/main/.github/workflows/tests_eessi_module.yml

ocaisa · 2026-04-17T12:13:56Z

bot: build repo:eessi.io-2025.06-software instance:eessi-bot-mc-aws for:arch=x86_64/amd/zen4

eessi-bot-aws · 2026-04-17T12:14:03Z

New job on instance eessi-bot-mc-aws for repository eessi.io-2025.06-software
Building on: amd-zen4
Building for: x86_64/amd/zen4
Job dir: /project/def-users/SHARED/jobs/2026.04/pr_206/148714

date	job status	comment
Apr 17 12:14:02 UTC 2026	submitted	job id `148714` awaits release by job manager
Apr 17 12:15:06 UTC 2026	released	job awaits launch by Slurm scheduler
Apr 17 12:20:16 UTC 2026	running	job `148714` is running
Apr 17 12:26:32 UTC 2026	finished	😢 FAILURE (click triangle for details) Details ✅ job output file `slurm-148714.out` ✅ no message matching `FATAL:` ❌ found message matching `ERROR:` ❌ found message matching `FAILED:` ❌ found message matching `required modules missing:` ❌ no message matching `No missing installations` ✅ found message matching `.tar.* created!` Artefacts `eessi-2025.06-software-linux-x86_64-amd-zen4-17764286740.tar.zst` size: 1 MiB (1604937 bytes) entries: 169 modules under 2025.06/software/linux/x86_64/amd/zen4/modules/all `EESSI-extend/2025.06-easybuild.lua` `elfutils/0.193-GCCcore-14.2.0.lua` software under 2025.06/software/linux/x86_64/amd/zen4/software `EESSI-extend/2025.06-easybuild` `elfutils/0.193-GCCcore-14.2.0` reprod directories under 2025.06/software/linux/x86_64/amd/zen4/reprod `elfutils/0.193-GCCcore-14.2.0/20260417_122321UTC` other under 2025.06/software/linux/x86_64/amd/zen4 no other files in tarball
Apr 17 12:26:32 UTC 2026	test result	😁 SUCCESS (click triangle for details) ReFrame Summary [ OK ] (1/5) EESSI_LAMMPS_lj %device_type=cpu %module_name=LAMMPS/22Jul2025-foss-2024a-kokkos %scale=1_node /ade8cad7 @BotBuildTests:x86-64-zen4+default P: perf: 1148.362 timesteps/s (r:0, l:None, u:None) [ OK ] (2/5) EESSI_OSU_coll %benchmark_info=mpi.collective.osu_allreduce %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node %device_type=cpu /e4bf9965 @BotBuildTests:x86-64-zen4+default P: latency: 1.45 us (r:0, l:None, u:None) [ OK ] (3/5) EESSI_OSU_coll %benchmark_info=mpi.collective.osu_alltoall %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node %device_type=cpu /3da4890b @BotBuildTests:x86-64-zen4+default P: latency: 3.26 us (r:0, l:None, u:None) [ OK ] (4/5) EESSI_OSU_pt2pt_CPU %benchmark_info=mpi.pt2pt.osu_latency %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node /3255009a @BotBuildTests:x86-64-zen4+default P: latency: 0.16 us (r:0, l:None, u:None) [ OK ] (5/5) EESSI_OSU_pt2pt_CPU %benchmark_info=mpi.pt2pt.osu_bw %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node /59f4b331 @BotBuildTests:x86-64-zen4+default P: bandwidth: 14296.92 MB/s (r:0, l:None, u:None) [ PASSED ] Ran 5/5 test case(s) from 5 check(s) (0 failure(s), 0 skipped, 0 aborted) Details ✅ job output file `slurm-148714.out` ❌ found message matching `ERROR:` ✅ no message matching `[\sFAILED\s].Ran . test case`

ocaisa · 2026-04-17T12:51:38Z

@zerefwayne

== FAILED: Installation ended unsuccessfully: Expected amdgcn_capabilities to be
set to build this EasyConfig. Please specify either --amdgcn-capabilities, or
set amdgcn_capabilities in the EasyConfig! (took 58 secs)

ocaisa · 2026-04-17T12:56:34Z

Wrong build command!

bot: build repo:eessi.io-2025.06-software instance:eessi-bot-mc-aws on:arch=zen4 for:arch=x86_64/amd/zen4,accel=amd/gfx90a

eessi-bot-aws · 2026-04-17T12:56:40Z

New job on instance eessi-bot-mc-aws for repository eessi.io-2025.06-software
Building on: amd-zen4
Building for: x86_64/amd/zen4 and accelerator amd/gfx90a
Job dir: /project/def-users/SHARED/jobs/2026.04/pr_206/148715

date	job status	comment
Apr 17 12:56:39 UTC 2026	submitted	job id `148715` awaits release by job manager
Apr 17 12:56:56 UTC 2026	released	job awaits launch by Slurm scheduler
Apr 17 12:57:58 UTC 2026	running	job `148715` is running
Apr 17 12:58:59 UTC 2026	finished	😢 FAILURE (click triangle for details) Details ✅ job output file `slurm-148715.out` ✅ no message matching `FATAL:` ❌ found message matching `ERROR:` ✅ no message matching `FAILED:` ✅ no message matching `required modules missing:` ❌ no message matching `No missing installations` ✅ found message matching `.tar.* created!` Artefacts `eessi-2025.06-software-linux-x86_64-amd-zen4-accel-amd-gfx90a-17764306630.tar.zst` size: 0 MiB (25379 bytes) entries: 13 modules under 2025.06/software/linux/x86_64/amd/zen4/accel/amd/gfx90a/modules/all no module files in tarball software under 2025.06/software/linux/x86_64/amd/zen4/accel/amd/gfx90a/software no software packages in tarball reprod directories under 2025.06/software/linux/x86_64/amd/zen4/accel/amd/gfx90a/reprod no reprod directories in tarball other under 2025.06/software/linux/x86_64/amd/zen4/accel/amd/gfx90a `2025.06/software/linux/x86_64/amd/zen4/modules/all/EESSI-extend/2025.06-easybuild.lua` `2025.06/software/linux/x86_64/amd/zen4/modules/devel/EESSI-extend/2025.06-easybuild.lua` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/easybuild-EESSI-extend-2025.06-20260417.125730.log` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/easybuild-EESSI-extend-2025.06-20260417.125730_test_report.md` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/EESSI-extend-2025.06-easybuild-easybuild-devel` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/EESSI-extend-2025.06-easybuild.eb` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/reprod/` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/reprod/easyblocks/` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/reprod/easyblocks/bundle.py` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/reprod/EESSI-extend-2025.06-easybuild.eb` `2025.06/software/linux/x86_64/amd/zen4/software/EESSI-extend/2025.06-easybuild/easybuild/reprod/EESSI-extend-2025.06-easybuild.env`
Apr 17 12:58:59 UTC 2026	test result	😁 SUCCESS (click triangle for details) ReFrame Summary [ OK ] (1/5) EESSI_LAMMPS_lj %device_type=cpu %module_name=LAMMPS/22Jul2025-foss-2024a-kokkos %scale=1_node /ade8cad7 @BotBuildTests:x86-64-zen4+default P: perf: 1059.685 timesteps/s (r:0, l:None, u:None) [ OK ] (2/5) EESSI_OSU_coll %benchmark_info=mpi.collective.osu_allreduce %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node %device_type=cpu /e4bf9965 @BotBuildTests:x86-64-zen4+default P: latency: 1.47 us (r:0, l:None, u:None) [ OK ] (3/5) EESSI_OSU_coll %benchmark_info=mpi.collective.osu_alltoall %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node %device_type=cpu /3da4890b @BotBuildTests:x86-64-zen4+default P: latency: 3.41 us (r:0, l:None, u:None) [ OK ] (4/5) EESSI_OSU_pt2pt_CPU %benchmark_info=mpi.pt2pt.osu_latency %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node /3255009a @BotBuildTests:x86-64-zen4+default P: latency: 0.15 us (r:0, l:None, u:None) [ OK ] (5/5) EESSI_OSU_pt2pt_CPU %benchmark_info=mpi.pt2pt.osu_bw %module_name=OSU-Micro-Benchmarks/7.5-gompi-2025a %scale=1_node /59f4b331 @BotBuildTests:x86-64-zen4+default P: bandwidth: 14282.18 MB/s (r:0, l:None, u:None) [ PASSED ] Ran 5/5 test case(s) from 5 check(s) (0 failure(s), 0 skipped, 0 aborted) Details ✅ job output file `slurm-148715.out` ❌ found message matching `ERROR:` ✅ no message matching `[\sFAILED\s].Ran . test case`

ocaisa · 2026-04-17T13:04:33Z

ERROR: Failed to parse configuration options: Found 1 environment variable(s) that are prefixed with EASYBUILD but do not match valid option(s): EASYBUILD_AMDGCN_COMPUTE_CAPABILITIES

ocaisa · 2026-04-17T13:07:04Z

bot: build repo:eessi.io-2025.06-software instance:eessi-bot-mc-aws on:arch=zen4 for:arch=x86_64/amd/zen4,accel=amd/gfx90a

eessi-bot-aws · 2026-04-17T13:07:11Z

New job on instance eessi-bot-mc-aws for repository eessi.io-2025.06-software
Building on: amd-zen4
Building for: x86_64/amd/zen4 and accelerator amd/gfx90a
Job dir: /project/def-users/SHARED/jobs/2026.04/pr_206/148716

date	job status	comment
Apr 17 13:07:09 UTC 2026	submitted	job id `148716` awaits release by job manager
Apr 17 13:08:03 UTC 2026	released	job awaits launch by Slurm scheduler
Apr 17 13:09:06 UTC 2026	running	job `148716` is running

ocaisa · 2026-04-17T13:18:38Z

We will also need to add a check to the EasyBuild hooks for software that requires an AMD GPU:

software-layer-scripts/eb_hooks.py

Lines 714 to 738 in 0aaadf7

    
           def pre_fetch_hook_check_installation_path(self, *args, **kwargs): 
        
               # When we know the CUDA status, we will need to verify the installation path 
        
               # if we are doing an EESSI or host_injections installation 
        
               accelerator_deps = ['CUDA'] 
        
               strict_eessi_installation = ( 
        
                   bool(re.search(EESSI_INSTALLATION_REGEX, self.installdir)) or 
        
                   self.installdir.startswith(HOST_INJECTIONS_LOCATION)) 
        
               if strict_eessi_installation and not os.getenv("EESSI_OVERRIDE_STRICT_INSTALLPATH_CHECK"): 
        
                   dependency_names = self.cfg.dependency_names() 
        
                   if self.cfg.name in accelerator_deps or any(dep in dependency_names for dep in accelerator_deps): 
        
                       # Make sure the path is an accelerator location 
        
                       if "/accel/" not in self.installdir: 
        
                           raise EasyBuildError( 
        
                               f"It seems you are trying to install an accelerator package {self.cfg.name} into a " 
        
                               f"non-accelerator location {self.installdir}. You need to reconfigure your installation to target " 
        
                               "the correct location." 
        
                               ) 
        
                   else: 
        
                       # If we don't have an accelerator dependency then we should be in a CPU installation path 
        
                       if "/accel/" in self.installdir: 
        
                           raise EasyBuildError( 
        
                               f"It seems you are trying to install a CPU-only package {self.cfg.name} into accelerator location " 
        
                               f"{self.installdir}. If this is a dependency of the package you are really interested in you will " 
        
                               "need to first install the CPU-only dependencies of that package." 
        
                               )

but this can be in a follow-up PR. Should be possible by taking the full list of dependencies (including toolchain deps which I think you can get from self.cfg._toolchain.tcdeps), and then just add adding ROCm-LLVM to acclerator_deps

Add AMD compute capabilities env if detected

7a0f948

zerefwayne force-pushed the eessi-extend-amd branch from cd8e08e to 7a0f948 Compare April 17, 2026 11:49

ocaisa requested changes Apr 17, 2026

View reviewed changes

ocaisa reviewed Apr 17, 2026

View reviewed changes

Comment thread easystacks/software.eessi.io/2025.06/accel/amd/eessi-2025.06-eb-5.3.0-001-system.yml Outdated

(temporary): add ROCm-LLVM to easystack

d1a699d

zerefwayne force-pushed the eessi-extend-amd branch from 4bc20c1 to d1a699d Compare April 17, 2026 12:10

ocaisa reviewed Apr 17, 2026

View reviewed changes

Comment thread EESSI-extend-easybuild.eb Outdated

Apply suggestion from @ocaisa

ec8e7e4

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add configuration for AMD to EESSI-extend-easybuild.eb#206

Add configuration for AMD to EESSI-extend-easybuild.eb#206
zerefwayne wants to merge 3 commits intoEESSI:mainfrom
zerefwayne:eessi-extend-amd

zerefwayne commented Apr 17, 2026

Uh oh!

ocaisa left a comment

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

eessi-bot-aws bot commented Apr 17, 2026 •

edited

Loading

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

eessi-bot-aws bot commented Apr 17, 2026 •

edited

Loading

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

eessi-bot-aws bot commented Apr 17, 2026 •

edited

Loading

Uh oh!

ocaisa commented Apr 17, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

zerefwayne commented Apr 17, 2026

Uh oh!

ocaisa left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

eessi-bot-aws bot commented Apr 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

eessi-bot-aws bot commented Apr 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026

Uh oh!

eessi-bot-aws bot commented Apr 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ocaisa commented Apr 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

eessi-bot-aws bot commented Apr 17, 2026 •

edited

Loading

eessi-bot-aws bot commented Apr 17, 2026 •

edited

Loading

eessi-bot-aws bot commented Apr 17, 2026 •

edited

Loading

ocaisa commented Apr 17, 2026 •

edited

Loading