SANTA CLARA, Calif., March 09, 2021 – Nvidia today announced AI Enterprise, a software suite of enterprise-grade AI development tools and frameworks supported by Nvidia with VMware vSphere 7 Update 2.
Nvidia said the software suite enables vSphere compute virtualization users to build AI applications with the same tools they use to manage data centers and hybrid cloud environments. The software “provides scale-out, multi-node, AI application performance on vSphere that is indistinguishable from bare-metal servers,” the company said in its announcement.
To power the applications in Nvidia AI Enterprise, VMware vSphere 7 Update 2 is now certified for Nvidia A100 Tensor Core GPUs on Nvidia-Certified Systems, which include servers from Dell Technologies, HPE, Lenovo and Supermicro.
“Until now, the world has run AI on bare-metal servers,” said Justin Boitano, vice president and general manager of Enterprise and Edge Computing at Nvidia. “Nvidia AI Enterprise enables customers to reduce AI model development time from 80 weeks to just eight weeks, and allows them to deploy and manage advanced AI applications on VMware vSphere with the same scale-out, record-breaking NVIDIA accelerated computing performance that’s possible on bare metal.”
“Every enterprise is exploring how to modernize their infrastructure to meet the demands of AI applications,” said Lee Caswell, vice president of marketing for the Cloud Platform Business Unit at VMware. “With NVIDIA AI Enterprise and vSphere 7 Update 2, VMware customers are now able to fast-track AI in their virtualized data centers and easily deploy certified, AI-ready infrastructure for their modern apps.”
Nvidia AI Enterprise provides compatibility for a broad set of accelerated CUDA applications, AI frameworks, pre-trained models and software development kits running in the hybrid cloud. Optimizations enable workloads to scale across multiple nodes to support large deep learning training models with full GPU virtualization.
Nvidia has also certified vSphere as the only compute virtualization software to provide hypervisor support for live migration with Nvidia Multi-Instance GPU technology, which allows each A100 GPU to be partitioned into up to seven instances at the hardware level to maximize efficiency for workloads of all sizes.
Additionally, select Nvidia ConnectX adapters are now certified for VMware vSAN over RDMA (Remote Direct Memory Access), which offloads CPU communication tasks to boost application performance and improve infrastructure returns on investment.
Nvidia AI Enterprise is available as a perpetual license at $3,595 per CPU socket. Enterprise Business Standard Support for Nvidia AI Enterprise is $899 annually per license. Customers can apply for early access to Nvidia AI Enterprise as they plan their upgrades to VMware vSphere 7 Update 2
