The demand for compute-intensive workloads, such as artificial intelligence, deep learning, and data science continues to grow and many customers are looking for flexible GPU solutions in order to dip into this realm.
GPU`s have been typically deployed as custom hardware cards which are installed inside of supported servers. NVIDIA has announced the release of the Virtual Compute Server (vCS).
The NVIDIA Virtual Compute Server (vCS) enables data centers to accelerate server virtualization with the latest NVIDIA data center GPUs, including NVIDIA A100 Tensor Core GPU, so that the most compute-intensive workloads, such as artificial intelligence, deep learning, and data science, can be run in a virtual machine (VM).
Support for the DellEMC VxRAIL platform has been added to their hardware compatibility matrix:
In regards to hypervisors: Red Hat RHV/RHEL supports vCS with the NVIDIA A100 Tensor Core GPU starting with the vGPU September 2020 release (11.1). Support for VMware vSphere will be available in an upcoming vSphere release (more to come on this later)
The ability to be able to run a GPU as a virtual appliance gives a lot of flexibility for customers that want all the features that virtualization has to offer, and can also provide the basis for an entry level solution for engineers that want to get their hands dirty without the need to buy dedicated GPU cards.
Also in regards to containers, NVIDIA says "Yes, containers can be run in VMs with vCS. NVIDIA NGC offers a comprehensive catalog of GPU-accelerated containers for deep learning, machine learning, and HPC. Workloads can also be run directly in a VM, without containers, using vCS".
You can read the full release briefing here: