Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


Overview of the HGX-1 AI Accelerator Chassis

Siamak Tavallaei

In this video from the Open Compute Summit, Siamak Tavallaei from Microsoft presents an overview of the Microsoft Project Olympus AI Accelerator Chassis, also known as the HGX-1.

“The Project Olympus hyperscale GPU accelerator chassis for AI, also referred to as HGX-1, is designed to support eight of the latest “Pascal” generation NVIDIA GPUs and NVIDIA’s NVLink high speed multi-GPU interconnect technology, and provides high bandwidth interconnectivity for up to 32 GPUs by connecting four HGX-1 together. The HGX-1 AI accelerator provides extreme performance scalability to meet the demanding requirements of fast growing machine learning workloads, and its unique design allows it to be easily adopted into existing datacenters around the world.”

Siamak Tavallaei is a Principal Architect at Microsoft’s Cloud & Enterprise division.  Collaborating with industry partners, he drives a number of initiatives in research, design, and deployment of hardware for Microsoft’s cloud-scale services such as Azure, Bing, Office 365, Exchange, and SQL across a global datacenter footprint With over 26 years of computer industry experience, he has been instrumental in development and evolution of innovative multi-processor servers and technology initiatives in areas of storage and memory hierarchy as well as heterogeneous, distributed computing. His server design and computer architecture journey started with Compaq in 1990.

Tavallaei designed ASICs and contributed to a number of chipsets, I/O buses (EISA, PCI, InfiniBand) and multi-processor bus protocol, memory, and storage initiatives within ProLiant server family and helped evolve industry-standard servers.  He held the rank of Principal Member Technical Staff at Compaq and was a Distinguished Technologist at Hewlett-Packard before joining Microsoft as a Principal Architect. Siamak holds over 30 patents in computer server architecture, design, and fault management; he has led and served on industry-wide initiatives; and he often shares his learning through published papers and presentations at industry conferences.

Download the Slides (PDF)

Sign up for our insideHPC Newsletter

Leave a Comment

*

Resource Links: