Container |
Cloud |
GPU |
Dockerfile |
ghcr.io/converged-computing/metric-lammps-gpu:azure-hpc-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-lammps-gpu:azure-hpc-reax |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-amg:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-kripke-gpu:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-laghos:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-minife:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-mixbench:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/mt-gemm:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-magma:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/multi-gpu-models:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-osu-gpu:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-nek5000:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-quicksilver-gpu:azure-hpc-gpu-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/metric-stream:azure-hpc-ubuntu2204 |
Azure |
yes |
Dockerfile |
ghcr.io/converged-computing/pytorch-resnet-experiment:azure-gpu |
Azure |
yes |
Dockerfile |
Since we need to vary builds across clouds, let's keep track of that here. These only include the ones we are intending to run.
Container |
Cloud |
GPU |
Dockerfile |
Notes |
ghcr.io/converged-computing/metric-amg2023:google-gpu |
Google |
yes |
Dockerfile |
This was GKE |
ghcr.io/converged-computing/metric-amg2023:spack-older-intel |
Google |
yes |
Dockerfile |
This was for smaller problem on compute engine |
ghcr.io/converged-computing/metric-amg2023:skylake-manual |
Google |
yes |
Dockerfile |
This was for larger problem on compute engine |
ghcr.io/converged-computing/metric-laghos:gpu-zen4 |
Google |
yes |
Dockerfile |
Needs to be built on large machine |
ghcr.io/converged-computing/metric-single-node:cpu-zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-kripke-gpu:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-gpu:kokkos |
Google |
yes |
Dockerfile |
using Kokkos build |
ghcr.io/converged-computing/metric-lammps-gpu:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-magma:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-minife:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-mixbench:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/mt-gemm:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/multi-gpu-models:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-nek5000:latest |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-osu-gpu:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-quicksilver-gpu:google-gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/pytorch-resnet-experiment:gpu |
Google |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-stream:google-gpu |
Google |
yes |
Dockerfile |
|
Container |
Cloud |
GPU |
Dockerfile |
Notes |
ghcr.io/converged-computing/metric-amg2023:spack-slim |
AWS |
yes |
Dockerfile |
Same as Google, already has libfabric |
ghcr.io/converged-computing/metric-laghos:libfabric-gpu-zen4 |
yes |
Dockerfile |
Needs to be built on large machine |
|
ghcr.io/converged-computing/metric-single-node:cpu-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-gpu:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-gpu:libfabric-reax |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-kripke-gpu:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-magma:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-minife:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-mixbench:libfabric-gpu |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/mt-gemm:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/multi-gpu-models:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-nek5000:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-osu-gpu:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-quicksilver-gpu:libfabric |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/pytorch-resnet-experiment:libfabric-gpu |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-stream:libfabric-zen4 |
AWS |
yes |
Dockerfile |
|
Note that Azure is different in that we use amg instead of amg2023. The base images are built based on the azhpc-images repository.
Container |
Cloud |
GPU |
Dockerfile |
ghcr.io/converged-computing/metric-lammps-cpu:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-kripke-cpu:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-amg2023:azure-hpc-cpu-int64-zen3 |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-laghos:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-lammps-cpu:azure-hpc-reax |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-single-node:cpu-zen4 |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-minife:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-nek5000:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-mixbench:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/mt-gemm:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-osu-cpu:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-quicksilver-cpu:azure-hpc |
Azure |
no |
Dockerfile |
ghcr.io/converged-computing/metric-stream:azure-hpc |
Azure |
no |
Dockerfile |
Container |
Cloud |
GPU |
Dockerfile |
Notes |
ghcr.io/converged-computing/metric-amg2023:spack-slim-cpu-int64-zen3 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-amg2023:rocky8-cpu-int64-zen3 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-laghos:cpu-zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-laghos:rocky-8 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-kripke-cpu:zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-kripke-cpu:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-lammps-cpu:zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-cpu:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-lammps-cpu:zen4-reax |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-cpu:rocky-8-reax |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-minife:cpu-zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-minife:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/mt-gemm:cpu-zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/mt-gemm:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-osu-cpu:zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-osu-cpu:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-quicksilver-cpu:zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-quicksilver-cpu:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-single-node:cpu-zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-stream:cpu-zen4 |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-stream:rocky-8 |
Google |
no |
Dockerfile |
Rocky Linux 8 |
ghcr.io/converged-computing/metric-mixbench:cpu |
Google |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-nek5000:cpu |
Google |
no |
Dockerfile |
|
For the above, Rocky Linux 8 are generally for the Terraform setup, and the others for GKE.
Container |
Cloud |
GPU |
Dockerfile |
Notes |
ghcr.io/converged-computing/metric-amg2023:spack-slim-cpu-int64-zen3 |
AWS |
no |
Dockerfile |
Same as Google, already has |
libfabric |
|
|
|
|
ghcr.io/converged-computing/metric-laghos:libfabric-cpu-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-single-node:cpu-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-kripke-cpu:libfabric-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-minife:libfabric-cpu-zen4 |
AWS |
yes |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-cpu:zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-lammps-cpu:libfabric-zen4-reax |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-mixbench:libfabric-cpu |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/mt-gemm:libfabric-cpu-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-osu-cpu:libfabric-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-quicksilver-cpu:libfabric-zen4 |
AWS |
no |
Dockerfile |
|
ghcr.io/converged-computing/metric-stream:libfabric-cpu-zen4 |
AWS |
no |
Dockerfile |
|
The AWS images are based off of the original Google cloud, but have oras and libfabric added. The CPU variants are the same as the GPU, with CPU stuffs removed. The exception is resnet, which uses the same pytorch base.