RYAX 23.06 aka "HPC release" is out !
We're happy to share the latest version of Ryax, with new amazing features, all dedicated to HPC - High Performance Computing ! As always, we're community-driven, and these features were asked by the community and the customers.
New features as
- HPC offloading using Singularity with multi user support
- Cuda GPU supports with the python3-cuda language
- Resources request support (CPU, Memory, Time, GPU)
- Keep the home .cache directory between runs
- Allows user to rebuild already built actions
- Better internal runs state management
- Use the latest Minio version
With HPC offloading in Ryax, you allow scientists to run their datascience apps on HPC without any change on their code. Security using Singularity, Monitoring and provisionning using slurm will ensure simplicity, and manageability.
With resources request support, you can define, for each action in a workflow, what are exactly the resources that will be used : number of CPUs, Memory size, time before the action is killed, number of GPUs ...), meaning, you can control exactly the resources used at each step, and optimise your HPC workload.
You're running an HPC center ?
and discover how Ryax helps you
to extend the usage of your HPC resources