...
| Note | |||||
|---|---|---|---|---|---|
Here we don't provide a complete list of available software. To check, what is currently available, please also check the output of
|
Compiler & MPI
We currently provide the following compilers:
| Name | Version | module | Notes |
|---|---|---|---|
| gcc | 8.5.0 | - | system default |
| 12.1.0 | gcc/12.1.0 | activated support for offloading on Nvidia GPUs (nvptx) | |
| intel-oneapi-compilers | 2022.1.0 | intel-oneapi-compilers/2022.1.0 | |
| nvhpc | 22.3 | nvhpc/22.3 | NVIDIA HPC Software Development Kit (SDK) |
| aocc | 3.2.0 | aocc/3.2.0 |
MPI
| AMD Optimizing C/C++ and Fortran Compilers (“AOCC”) | |||
| Name | Version | module | Notes |
|---|---|---|---|
| openmpi | 4.1.3 | openmpi/4.1.3 | |
| intel-oneapi-mpi | 2021.6.0 | intel-oneapi-mpi/2021.6.0 |
Compiler options for optimization
Some remarks from Natalja with respect to the intel compiler and the FESOM2 benchmark (used by NEC in there albedo offer):
- do not use -xHost, because Intel does not "recognize" AMD (officially for security reasons
. Therefore use: -xcore-avx2 - These options were used by NEC during the FESOM2 benchmark:
PT = -O3 -qopt-report5 -no-prec-div -ip -fp-model=fast=2 -implicitnone -march=core-avx2 -fPIC –qopenmp -qopt-malloc-options=2 -qopt-prefetch=5 -unroll-aggressiveThese are (at least partially) quite important for good performance. However, we do not have the experience which are more or less critical. Be careful, some options might kill reproducibility (e.g., -fp-model=fast=2).
Spack
On albedo we mainly use spack to install software and provide module files.
...