Connect with us

Computing

Habana Labs AI Accelerators To Be Used in Voyager Supercomputer

Published

 on

The San Diego Supercomputer Center (SDSC) at UC San Diego has selected Habana Labs’ artificial intelligence (AI) training and inference accelerators to provide high-performance, high-efficiency AI computing in its Voyager supercomputer. Habana Labs is a leading developer of purpose-built deep learning accelerators.

Voyager Supercomputer

The Voyager supercomputer will scale AI training capacity with Habana’s unique interconnect technology, specifically 336 Gaudi processors. These processors are especially good for scaling large supercomputer training applications, and Gaudi is the only AI processor in the industry that can natively integrate ten 100-Gigabit Ethernet ports of RoCE v2 on-chip. This enables flexibility of scaling and helps avoid throughput bottlenecks, which can limit scaling capacity. Besides this, the Voyager system also contains 16 Habana Goya processors, which are used to power AI inference models. 

Amitava Majumdar is head of SDSC’s Data Enables Scientific Computing division and principal investigator for the Voyager project. 

“With innovative solutions optimized for deep learning operations and AI workloads, Habana accelerators are ideal choices to power Voyager’s forthcoming AI research,” Majumdar said. “We look forward to partnering with Habana, Intel and Supermicro to bring this uniquely efficient class of compute capabilities to the Voyager program, giving academic researchers access to one of the most capable AI-focused systems available today.”

Performance and efficiency in AI operations are increased with Habana Gaudi AI training and Goya inference processors. For data scientists and researchers that have access to Voyager, they will be able to customize models with programmable Tensor Processor Cores and kernel libraries. They will also be able to ease implementation by relying on Habana’s SynapseAI software platform. The platform supports popular machine learning frameworks and AI models for applications like vision, natural language processing and recommendation systems. 

AI Systems Provided by Supermicro

Habana-based AI systems will be provided for Voyager by Supermicro, a global leader in enterprise computing, storage, networking and green computing technology, as well as a Voyager technology partner. 

The systems include:

  • Supermicro X12 Gaudi AI Training System – eight Gaudi HL-205 cards paired with dual-socket 3rd Gen Intel Xeon Scalable processors.
  • Supermicro SuperServer 4029GP-T featuring eight Goya HL-100 PCLe cards for AI inference, paired with dual-socket 2nd Gen Intel Xeon scalable processors. 

Ray Pang is vice president of technology and business enablement at Supermicro. 

“Combining Supermicro’s advanced application-optimized server and storage hardware with Habana’s AI training and inference products is the best solution for SDSC’s multi-year Voyager AI project,” said Pang. “We continue to work closely with leading innovators to deliver solutions for computationally intensive projects worldwide at leading research and HPC environments for science and medical discovery, compute, and leading-edge AI solutions.”

Eitan Medina is Chief Business Officer at Habana. 

“We are honored that Habana’s AI processors have been selected to power the AI workloads that will run on San Diego Supercomputer Center’s Voyager supercomputer,” said Medina. “This implementation of our Gaudi and Goya products showcases how top academic institutions like SDSC can harness efficiency and performance to effectively address the growing demands of AI research workloads.”

The system will be in the Testbed Phase for the first three years of operation, and SDSC will work with research teams from astronomy, climate sciences, chemistry, particle physics and other fields. During this testing phase, SDSC will collaborate with the AI research community and document processes that could be used as a resource in the future. 

Trish Damkroger is vice president and general manager of Intel’s High Performance Computing group.

“The level of performance and efficiency that Voyager will require is precisely what Intel architectures are designed for,” said Damkroger. “Our Xeon Scalable processors coupled with Habana AI accelerators will ensure Voyager’s users have the HPC and AI capabilities they need to power their game-changing research.”

 

Alex McFarland is a historian and journalist covering the newest developments in artificial intelligence.