TodayMarch 20, 2023
The demand for advanced AI continues to grow. The answer today is a supercomputing infrastructure, services and expertise to accommodate the exponentially increasing size and complexity of the latest AI models.
For this reason, Microsoft is now introducing its most powerful and massively scalable VM series for AI applications on Azure: the ND H100 v5 VM enables on-demand in sizes from eight to thousands of NVIDIA H100 GPUs powered by NVIDIA Quantum-2 InfiniBand networks are interconnected. Azure customers get significantly better performance for AI models with technologies such as:
- 8x NVIDIA H100 Tensor Core GPUs connected via next gen NVSwitch and NVLink 4.0
- 400 Gb/s NVIDIA Quantum-2 CX7 InfiniBand per GPU with 3.2 TB/s per VM in a non-blocking fat tree network
- NVSwitch and NVLink 4.0 with 3.6 TB/s bi-sectional bandwidth between 8 local GPUs within each VM
- 4th Gen Intel Xeon Scalable processors
- PCIE Gen5 Host to GPU connected with 64 GB/s bandwidth per GPU
- 16 channels of 4800MHz DDR5 DIMMs
Exascale AI supercomputers for the cloud
Generative AI applications are evolving rapidly and offer unique value in almost every industry. By deploying exascale supercomputers, Microsoft customers around the world can achieve true supercomputing performance for their AI projects.