Test drive an Nvidia GPU in Supermicro’s STEP programme
Customers can test drive Supermicro’s 2U HGX A100 4-GPU or 4U HGX A100 8-GPU system with Nvidia third generation NVLink technology as part of a GPU programme initiated by Supermicro.
STEP (Supermicro Test drive Engagement with Partners) allows customers to remote test drivethe GPU systems remotely.
The collaboration with Nvidia will allows developers to “showcase the power of these servers to support unique applications and accelerate time-to-market solutions,” said Don Clegg, senior vice president, Worldwide Sales at Supermicro.
The programme can be accessed via the landing page with links to participating partners where customers can begin the process of registration. Once registered, customers can link directly to remote Supermicro Nvidia HGX A100 platforms to test and qualify workloads.
The Nvidia HGX AI supercomputing platform is purpose-built for the highest performance on simulation, data analytics and [artificial intelligence] AI applications, explained Nvidia’s senior director of Product Management and Marketing, Paresh Kharya.
Supermicro’s advanced high-density 2U and 4U servers feature Nvidia HGX A100 four- and eight-GPU baseboards. Supermicro’s Advanced I/O Module (AIOM) form factor is claimed to further enhance networking communication with high flexibility. The module can be coupled with the latest high speed, low latency PCI-E 4.0 storage and networking devices that support Nvidia GPUDirect RDMA and GPUDirect Storage with NVMe over Fabrics (NVMe-oF) on Nvidia Mellanox InfiniBand that feeds the scalable multi-GPU system with a continuous stream of data flow without bottlenecks.
Participating partners vary according to region. In EMEA, Boston, offers customers the opportunity to test one of Supermicro’s highest-performance servers based on Nvidia A100 GPUs, said managing director ,Manoj Nayee. “The Supermicro AS -2124GQ-NART system is an ideal building block for AI cluster requirements offered in the compact 2U form factor that provides an impressively dense configuration supporting four Nvidia A100 GPUs,” he explained. The team at Boston Labs and the Boston Training Academy looks forward to supporting customers for testing, training, and implementation of this next generation of accelerated computing,” he added.
Another partner is Broadberry Data Systems, which offers clients remote access to Nvidia A100 GPU-based servers, the Broadberry CyberServe A+ Server 2124GQ-NART and CyberServe EPYC EP2-4124GO-NART GPU server.
