Home Enterprise Supermicro Announces Redstone-based Server

Supermicro Announces Redstone-based Server

by Michael Rink

Today, Supermicro announced two new servers that will use NVIDIA’s recently teased third generation of HGX GPUs (Graphics Processing Units), NVIDIA A100 GPUs. SMCI, Super Micro Computer Incorporated, commonly referred to as just Supermicro, was founded in 1993 and is one of the fastest-growing IT companies in the world.

Today, Supermicro announced two new servers that will use NVIDIA’s recently teased third generation of HGX GPUs (Graphics Processing Units), NVIDIA A100 GPUs. SMCI, Super Micro Computer Incorporated, commonly referred to as just Supermicro, was founded in 1993 and is one of the fastest-growing IT companies in the world.

Supermicro NVIDIA A100

Unlike the other Supermicro servers we’ve covered recently, both of Supermicro’s upcoming new servers are targeted at the artificial intelligence (AI) and deep learning markets. These new servers are intended to be upgrades from Supermicro’s existing high-end AI-focused machines, which use the second generation NVIDIA HGX-2 Tesla V100 GPUs. The first of the new appliances is a 2U rack unit built around the new NVIDIA HGX A100 4 GPU board (formerly codenamed Redstone). The second appliance is twice the size. It is a 4U rack unit built around the new NVIDIA HGX A100 8 GPU board (previously codenamed Delta). Supermicro is also planning to add support for the NVIDIA A100 GPUs, when they are released, across the company’s broad portfolio of 1U, 2U, 4U, and 10U GPU servers.

Supermicro’s new 2U appliance has one PCI-E 4.0 x8 and up to four PCI-E 4.0 x16 expansion slots. Like previous HGX boards, all four of the GPUs will be interconnected through NVIDIA NVLink and capable of sharing up to 600GB per second.

Supermicro’s new 4U appliance has eight expansion slots for GPUDirect RDMA high-speed network cards and ten PCI-E 4.0 x16 expansion slots. The company projects that the eight GPUs, interconnected through NVIDIA NVSwitch, will have a combined speed possibly as high as 5 PetaFLOPS in just one rack unit. However, this news is so recent that the company hasn’t even shared the names or model numbers of the new servers yet, so these numbers may change.

Supermicro GPU Servers

Discuss on Reddit

Engage with StorageReview

Newsletter | YouTube | Podcast iTunes/Spotify | Instagram | Twitter | Facebook | RSS Feed