Spack Modules Release 22.1.0

Note: The software stack spack/release/22.1.0 is a stable rollout. The default will be Spack/release/21.1.1 and the release 22.1.0 is an additional stack with newer compilers and MPI for users to try out the latest tools. A number of additional packages along with newer Intel-mpi, intel-mkl, and compilers are provided in comparison to Release/21.1.1 stack.

Change Log

Changes

current state

previous (Spack Modules Staging 21.1.1)

previous (Spack Modules Release 21.1.1)
Available on Plattformx86_64, Haswell (CoolMUC2), Skylake (SuperMUC-NG)x86_64, Haswell (CoolMUC2)Skylake (SuperMUC-NG)
OSSLES 15.1SLES 15.1SLES 15.1
Spack Version0.16.20.16.10.16.1
Environment Modules Version

4.6.1

4.6.1

4.6.1

Compilers

Available versions are,

intel/18.0.5, intel/19.0.4, intel/19.0.5, intel/19.1.2, intel/20.0.4

gcc/8  gcc/8.5.0  gcc/9  gcc/9.4.0  gcc/10  gcc/10.3.0  gcc/11.2.0 

nvhpc/21.7

nag/7.0 

llvm/9.0.1  llvm/10.0.1  llvm/11.0.1



intel/19.0.5

gcc/8.4.0


intel/19.0.5

gcc/8.4.0

Module Names

PACKAGE/version-compiler&version-othersuffixs

Example:

fgsl/1.3.0-gcc10

fgsl/1.3.0-intel19

  • Generic installations for many packages are installed under the x86_64 tree.
    • These include compilers (gcc, pgi, nag, intel), tools (git, tmux,...), and graphics packages.
  • Application packages, performance libraries, and MPI parallel packages are installed in haswell, skylake-avx512, mic-knl trees.
  • The spack module will provide both the generic and the matching architecture-specific module tree.


MPI Module name

Available versions are,

intel-mpi/2019.12.320

intel-mpi/2018.4.274

  • Version 2018.x is provided but it is not supported by Intel on SLES15; use at own risk
  • Intel MPI Module with GCC compatibility is suffixed by gcc, 

intel-mpi/version-gcc

intel-mpi/2019.10.317

intel-mpi/2018.4.274

intel-mpi/2019.10.317

intel-mpi/2018.4.274

MKL Module name

Available versions are,

intel-mkl/2018.4.274

intel-mkl/2019.5.281

intel-mkl/2020.4.304

Note: specific linking variables can be obtained by providing module alias names with the corresponding suffixes
also the "fftw" interface can be activated this way.

intel-mkl/2018.4.274

intel-mkl/2019.5.281

intel-mkl/2020.1.217

intel-mkl/2018.4.274

intel-mkl/2019.5.281

intel-mkl/2020.1.217

Performance tools (advisor, inspector, itac, ...) Modules

Available versions are,

intel-parallel-studio/cluster.2018.4

intel-parallel-studio/cluster.2019.5

intel-parallel-studio/cluster.2020.4

Note: You will need to unload the modules intel-mkl, intel-mpi, intel (in this order) to use it



Modules Alias
  • Module alias names are used to provide default versions for a package or a major version.
  • They point to installed modules with full version numbering.
  • Use the "module aliases" command to get an overview.
  • Module alias names are used to provide certain variants of a software module,
  • e.g.
    • intel-mpi/2019-gcc provides the gcc, enables the compiler wrappers mpicc, mpif90,...
    • intel-mpi/2019-intel provides the icc, ifort enabled wrappers although both point to the same module.
  • Users can define their own alias names to get specific module provided variables,
  • e.g.
    • intel-mkl/2019-i8-gcc8 to obtain environment variables to link mkl with gcc-8 support and long integers


Python modules

Available versions are,

python/3.7.10-base 

python/3.7.10-extended 

python/3.8.8-base 

python/3.8.8-extended

Please check the python documentation.
The old intel python modules are still available but need to be loaded explicitly.



Current Modules: List