Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
67 commits
Select commit Hold shift + click to select a range
d0f7cbf
Added YAML files for 26.2/gh200
stelliom Mar 10, 2026
75ed935
Update config.yaml
stelliom Mar 10, 2026
645f68c
Quick test
stelliom Mar 18, 2026
e319f84
Remove pinned libtool version
stelliom Mar 18, 2026
2dcd6fd
Remove eccodes
stelliom Mar 18, 2026
6056608
Readd eccodes
stelliom Mar 18, 2026
4011ba0
Pin cuda@12.6.0
stelliom Mar 19, 2026
6729272
Changes based on old gh200 recipe
stelliom Mar 25, 2026
9ce44c3
Removed libtool pin
stelliom Mar 25, 2026
3793be5
Revert "Removed libtool pin"
stelliom Mar 25, 2026
2527c8c
Revert "Changes based on old gh200 recipe"
stelliom Mar 25, 2026
fcdb5f1
Use gcc 13
stelliom Mar 25, 2026
b1f168b
Revert to gcc 12.3.0
stelliom Mar 25, 2026
7d0cf52
Upgrade compiler versions
stelliom Mar 27, 2026
1aca090
Upgrade python and cuda versions
stelliom Mar 27, 2026
488cc9f
Revert gcc version
stelliom Mar 27, 2026
2377168
Upgrade commit of spack-packages to get cuda 13.1.1
stelliom Mar 27, 2026
3d65787
Reverted upgraded cuda version
stelliom Mar 27, 2026
709e16e
Upgrade to cuda 12.7.0
stelliom Mar 27, 2026
7bddd68
Use cuda 12.6.3
stelliom Mar 27, 2026
e1811a1
Use cuda 12.8.0
stelliom Mar 27, 2026
33834a1
Upgraded to latest versions of nvhpc, cuda and cray-mpich
stelliom Apr 16, 2026
dc4cf55
Try updating to Spack develop branch
stelliom Apr 16, 2026
06b44eb
Revert to nvhpc 26.1
stelliom Apr 16, 2026
e416591
Don't pin cray-mpich
stelliom Apr 16, 2026
2ffd35f
Specify compilers
stelliom Apr 18, 2026
e200c0e
Change unify to when_possible
stelliom Apr 18, 2026
95e12e9
Removed cxx from netcdf-fortran
stelliom Apr 18, 2026
eb51901
Removed cxx from hdf5
stelliom Apr 18, 2026
a757db9
Use nvhpc 26.3
stelliom Apr 28, 2026
9d322d7
Use cuda 13.1
stelliom Apr 28, 2026
7dc1ba1
Update eccodes and hdf5 to latest available version
stelliom May 5, 2026
18ef043
Revert hdf5 to 1.14.6
stelliom May 5, 2026
1be3568
Revert eccodes to 2.36.4 for consistency with eccodes-cosmo-resources
stelliom May 5, 2026
985d758
Remove pin for libtools
stelliom May 5, 2026
8d0aae2
Uncomment boost, nccl and extras
stelliom May 5, 2026
e46617d
Uncomment cdo
stelliom May 5, 2026
54ad2d2
Test gcc 14
stelliom May 11, 2026
0b9d6b5
Pin libfabric version as a dependency of cray-mpich
stelliom May 12, 2026
9109693
Reverted to GCC 13
stelliom May 12, 2026
8f8e6fc
Remove activation line from reframe.yaml
stelliom May 19, 2026
d4fd454
Added additional packages to reframe tests
stelliom May 19, 2026
0de05c1
Try unify true again
stelliom May 19, 2026
00fd07d
Revert to when_possible and add duplicates strategy
stelliom May 19, 2026
04b885d
Use gcc for libfabric
stelliom May 19, 2026
b2fdf33
Add empty specs for MPI
stelliom May 19, 2026
7646fd6
Remove compiler for libfabric
stelliom May 19, 2026
27c8125
Try overspecifying libfabric
stelliom May 19, 2026
e7c7e19
Reverted overspecification of libfabric
stelliom May 20, 2026
49ff865
Quick test
stelliom May 20, 2026
0ebb3fc
Add libfabric to the spec
stelliom May 20, 2026
91aa5e2
Move libfabric to mpi specs
stelliom May 20, 2026
e9bdc9b
Remove libfabric as a dependency
stelliom May 20, 2026
a814ea6
Reverted previous commit and pinned nccl
stelliom May 20, 2026
732d0b2
Quick test
stelliom May 20, 2026
ff1a17c
Specify compilers for all packages
stelliom May 20, 2026
2f611b0
Reverted last two commits
stelliom May 20, 2026
2e75d0f
Added xpmem to network specs
stelliom May 20, 2026
a6cee92
Temporarily remove aws-ofi-nccl
stelliom May 20, 2026
5a2eb2b
Removed xpmem from specs
stelliom May 21, 2026
3fac875
Use gcc for libfabric
stelliom May 21, 2026
e2fba2f
Readd xpmem and use nvhpc for libfabric
stelliom May 21, 2026
933524a
Remove NCCL
stelliom May 21, 2026
25f4b92
Readd NCCL
stelliom May 21, 2026
4f7d9d7
Use GCC 14
stelliom May 21, 2026
3a79933
Compile aws-ofi-nccl with nvhpc
stelliom May 21, 2026
cbef501
Revert previous commit
stelliom May 21, 2026
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions config.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -39,6 +39,11 @@ uenvs:
zen3: '5.18'
a100: '5.18'
icon:
"26.2":
recipes:
gh200: '26.2/gh200'
deploy:
santis: [gh200]
"25.2":
recipes:
gh200: '25.2/gh200'
Expand Down
4 changes: 4 additions & 0 deletions recipes/icon/26.2/gh200/compilers.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
gcc:
version: "14"
nvhpc:
version: "26.3"
11 changes: 11 additions & 0 deletions recipes/icon/26.2/gh200/config.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
name: icon
store: /user-environment
spack:
repo: https://github.com/spack/spack.git
commit: releases/v1.1
packages:
repo: https://github.com/spack/spack-packages.git
commit: develop
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just a reminder to pin this before finalizing the uenv. We typically do

Suggested change
commit: develop
commit: <sha> # develop on <date>

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you 🙏

modules: true
description: "Compilers and libraries required to build ICON"
version: 2
37 changes: 37 additions & 0 deletions recipes/icon/26.2/gh200/environments.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,37 @@
icon:
compiler: [gcc, nvhpc]
network:
mpi: cray-mpich@9.1.0 +cuda %c,cxx,fortran=nvhpc ^libfabric@2 +gdrcopy %c=nvhpc
specs: [ xpmem ]
unify: when_possible
specs:
- python@3.12 %c,cxx=gcc ^libffi %c,cxx=gcc ^expat %c=gcc ^libmd %c=gcc ^libxcrypt %c=gcc ^gettext %c,cxx=gcc
Comment thread
stelliom marked this conversation as resolved.
- cuda@13.1
- eccodes@2.36.4 +aec+fortran~ipo~memfs~netcdf~openmp~png~pthreads+shared~tools jp2k=openjpeg %c,cxx,fortran=nvhpc
- eccodes-cosmo-resources@2.36.0.3
- hdf5@1.14.6 ~cxx+fortran+hl~ipo~java~map+mpi+shared~subfiling+szip~threadsafe+tools api=default %c,fortran=nvhpc
- netcdf-c@4.9.3 +blosc~byterange~dap~fsync~hdf4~jna~logging+mpi~nczarr_zip+optimize~parallel-netcdf+pic+shared+szip+zstd %c,cxx=nvhpc
- netcdf-fortran@4.6.2 ~doc+pic+shared %c,fortran=nvhpc
- boost ~mpi
- openblas %c,cxx,fortran=nvhpc
# for validation
- cdo
# everything needed for nccl on SS11
- nccl@2.29.2
- nccl-tests
- aws-ofi-nccl@1.17.2
# extras
- libfyaml
- zlib-ng
- cmake
- findutils
variants:
- +mpi
- +cuda
- cuda_arch=90
views:
default:
link: roots
uenv:
prefix_paths:
LD_LIBRARY_PATH: [lib, lib64]
15 changes: 15 additions & 0 deletions recipes/icon/26.2/gh200/extra/reframe.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
default:
features:
- cuda
- mpi
- cray-mpich
- nccl
- nccl-tests
- openmp
- osu-micro-benchmarks
- serial
cc: mpicc
cxx: mpic++
ftn: mpifort
views:
- default
26 changes: 26 additions & 0 deletions recipes/icon/26.2/gh200/modules.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,26 @@
modules:
# Paths to check when creating modules for all module sets
prefix_inspections:
bin:
- PATH
lib:
- LD_LIBRARY_PATH
lib64:
- LD_LIBRARY_PATH

default:
arch_folder: false
# Where to install modules
roots:
tcl: /snap/modules
tcl:
all:
autoload: run
hash_length: 0
hide_implicits: true
exclude:
- '%gcc@7.5.0'
- '%gcc@12'
- 'gcc %gcc@7.5.0'
projections:
all: '{name}/{version}'