Skip to content

Releases: E4S-Project/e4s

E4S 24.05

06 May 19:07
2c1111e
Compare
Choose a tag to compare

E4S 24.05 includes a number of improvements from the prior quarterly release

  • Many packages are from prior release have been updated with newer versions
  • CUDA upgraded from CUDA 11 to CUDA 12 in most places
  • ROCm upgraded from ROCm 5.4.3 to ROCm 5.7.1
  • GROMACS builds added for NVIDIA GPU
  • New packages:
    • py-deephyper
    • nek5000
    • nekbone
    • e4s-alc
    • e4s-cl
    • fpm
    • netcdf-fortran
  • Model environment OS upgraded from Ubuntu 20.04 to Ubuntu 22.04
  • New Spack build cache containing specs for model environments: https://cache.e4s.io/24.05

E4S 24.02

14 Feb 23:34
Compare
Choose a tag to compare

E4S 24.02 includes a number of improvements from the prior quarterly release

  • Many packages are released with new versions
  • GROMACS molecular dynamics package has been added
  • CP2K quantum chemistry and solid state physics package has been added
  • Intel OneAPI MPI now used in model environment for Intel GPU
  • Intel OneAPI version upgraded to latest 2024 release
  • New Spack build cache containing specs for model environments https://cache.e4s.io/24.02

E4S 23.11

09 Nov 20:59
27db068
Compare
Choose a tag to compare

E4S 23.11 includes a number of improvements from the prior quarterly release

  • Many packages are released with new versions
  • ExaGO for large-scale power grid optimization code is now included in E4S, with builds capable of using both AMD and NVIDIA GPUs
  • Improved support for Intel GPUs:
    • PETSc with SYCL support added
    • SLATE with SYCL support added
    • SUNDIALS SYCL-enabled build errors resolved
  • New Spack build cache containing specs for model environments https://cache.e4s.io/23.11

E4S 23.08

31 Aug 19:54
Compare
Choose a tag to compare

E4S 23.08 includes a number of improvements from the prior quarterly release

  • Many packages are released with new versions
  • Improved support for NVIDIA Hopper GPU architecture
    • For ARM users: a number of packages have been added with support for NVIDIA Hopper architecture (cuda_arch=90)
    • A number of fixes for +cuda cuda_arch=90 packages for x86_64
  • Trilinos CPU version is increased to 14.4.0 where before it was 13.0.1
  • Libpressio CUDA build errors have been resolved
  • Improved ability to make use of binary cache with concretizer:reuse:true
  • Model environments designed to work with new series of base containers equipped with mpich
  • New Spack build cache containing specs for model environments https://cache.e4s.io/23.08
  • ROCm model environment contains example packages: configuration for using an externally installed ROCm stack