Spack

Installation

git clone https://github.com/spack/spack.git
cd spack
git checkout releases/latest
source ./share/spack/setup-env.sh

Please add the last command to .bashrc with absolute path to start spack by default.

Compiler

• Add Compilers

  spack install gcc@10.2.0 ## specify version
  spack find -p gcc ## get location
  spack compiler add /jet/packages/spack/opt/spack/linux-centos8-zen/gcc-8.3.1/gcc-10.2.0-tfzxq7udz2a53dmujvasy4uz33t27iwv ## or the location of any other compilers

Common Commands

The -j 32 flag in spack install means to make with 32 jobs, which will reduce installation time

spack info cuda # find available version of cuda
spack install -j 32 cuda@11.8 # install specified version of cuda
spack load cuda@11.8 # load corresponding cuda

Environment

• Create an Environment

  spack env create neko
  spack env activate neko

• Add Packages to the Environment Edit spack.yaml from spack/var/spack/environments/neko

  cd var/spack/environments/neko
  vim spack.yaml

  And modify the file to add the package you need.

  spack:
    specs:
    - zlib
    - binutils
    - cuda@11.8
    - numactl
    - xz
    - libxml2
    - libpciaccess
    - hwloc
    - libevent
    - check
    - gdrcopy
    - ucx
    - openmpi+cuda
    - openblas
    - pmix
    - fftw
    - scalapack
    - libfabric
    view: true
    concretizer:
      unify: true
  

  Or you can just type

  spack add cuda@11.8

  Concretize the packages:

  spack concretize

  Finally, install all packages in the environment

  spack install --overwrite

  More in Reference

yaml Setting

By adding a packages.yaml to the .spack folder, you can change the default compiler and use external cuda to build openmpi.

• The +cuda is used to compile openmpi with cuda support
• % means to specify the compiler used so that packages match with each others

packages:
  all:
    compiler: [gcc@12.3.0]
  openmpi:
    variants: +cuda
  cuda:
    externals:
    - spec: cuda@11.8%gcc@12.3.0
      prefix: /usr/local/cuda/
    buildable: false

Conda

Installation

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
sh Miniconda3-latest-Linux-x86_64.sh

Terminal needs to be restarted.

Create Environment

conda create -n my_env python=3.11 # or any other environment names and python
conda activate my_env
conda install numpy