scorec-spack-packages (Spack repo)

This Spack repository (namespace scorec) provides recipes for SCOREC software and related dependencies, including:

• PCMS: SCOREC/pcms
• MeshFields: SCOREC/meshFields
• EnGPar: SCOREC/EnGPar
• PUMI-PiC: SCOREC/pumi-pic
• M3D-C1: PrincetonUniversity/M3DC1
• OpenMC with PUMI‑Tally (OpenMC fork integrating PUMI‑Tally):
  • OpenMC fork: Fuad-HH/openmc (upstream: openmc-dev/openmc)
  • PUMI‑Tally: Fuad-HH/PumiUMTally

Installation

1. Install Spack

It is recommended to use the system-provided Spack if available. If not, follow the instructions below to install Spack from source. Check the Spack documentation for more details.

git clone --depth=2 https://github.com/spack/spack.git
. /your/path/to/spack/share/spack/setup-env.sh

2. Create a new environment and activate it

spack env create pumi-env
spack env activate pumi-env

3. Add this repository to Spack and update the builtin repository

spack repo add https://github.com/SCOREC/scorec-spack-packages.git --name scorec
spack repo update builtin
spack repo update builtin --branch releases/v2026.10

Now, if you do spack config get repos you should see the repositories added to the list. And doing spack info pumi-pic should show the package.

4. Install Packages

spack add pumi-pic

or for OpenMC with PUMI-Tally:

spack add openmc-pumi +pumitally +dagmc ^kokkos+openmp+serial

and run spack concretize and spack install to install the packages.

Note

PUMI-PiC also supports other Kokkos backends. Check available backends with spack info kokkos.

Using OpenMC with PUMI-Tally

The case setup and run instructions are available here: OpenMC with PUMI-Tally

Build cache and container images

CI builds the environments under environments/ and pushes the results to a public Spack build cache. Add it as a mirror to install prebuilt binaries instead of building from source:

spack mirror add --unsigned scorec oci://ghcr.io/scorec/scorec-spack-buildcache

PCMS container image

The pcms+python+petsc spec of environments/pcms is also published as a runnable, multi-architecture container image. See the PCMS README for how to use it.

Tag	Contents
ghcr.io/scorec/pcms:develop	Latest build, for linux/amd64 and linux/arm64
ghcr.io/scorec/pcms:develop-YYYY-MM-DD	The build from that date

The same images are also tagged in the build cache repository as pcms-develop, pcms-develop-YYYY-MM-DD, and single-arch containers pcms-develop-amd64 and pcms-develop-arm64.

M3D-C1 container image

environments/m3dc1-image is published as ghcr.io/scorec/m3dc1, for linux/amd64 (x86-64-v3) only. It contains these executables, all built with ADAS support, plus the SCOREC mesh utilities (split_smb, ...):

Executable	Features
m3dc1_2d, m3dc1_2d_complex, m3dc1_3d, m3dc1_3d_st	Regression-test flavors
m3dc1_2d_omp_pic, m3dc1_2d_complex_omp_pic, m3dc1_3d_omp_pic, m3dc1_3d_omp_st_pic	Same, with OpenMP and particles

The OpenADAS ADF11 data is included. Set the adas_adf11 input in C1input to the value of $M3DC1_ADAS_ADF11.

# The container runs as root, which Open MPI only allows when told to.
docker run --rm -it -v "$PWD:/work" -w /work \
  -e OMPI_ALLOW_RUN_AS_ROOT=1 -e OMPI_ALLOW_RUN_AS_ROOT_CONFIRM=1 \
  ghcr.io/scorec/m3dc1:develop mpirun -n 4 m3dc1_2d

Tag	Contents
ghcr.io/scorec/m3dc1:develop	Latest build
ghcr.io/scorec/m3dc1:develop-YYYY-MM-DD	The build from that date