Supermicro (NASDAQ: SMCI), a global provider of compute, storage, and networking technologies, has unveiled its new BigTwin server architecture. The BigTwin is the fourth generation of Supermicro’s Twin architecture that would provide massively improved performance-per-watt, per-dollar, and per-square-foot as well as hot-swap U.2 NVMe.
The BigTwin is a 2U rackmount server available in either two or four node models. Each node supports all of the following: 24 DIMM of ECC DDR4-2400MHz and higher for up to 3TB of memory; flexible networking with SIOM add-on cards with quad/dual 1GbE, quad/dual 10GbE/10G SFP+, dual 25G, 100G, FDR and EDR InfiniBand options; 6 hot-swap 2.5″ NVMe / SAS3 / SATA3 drives; 2 PCI-E 3.0 x16 slots; M.2 and SATADOM; high-efficiency, high-density power stick technology; and DP Intel Xeon processor E5-2600 v4/v3 product families up to 165W and 200W.
Sold as a complete system for “highest product quality, delivery, and performance,” the BigTwin is supported by Supermicro IPMI software and Global Services, and is optimized for HPC, data center, cloud, and enterprise environments.
“Innovation is at the core of Supermicro product development and benefits the HPC community with first-to-market integration of advanced technology such as our 1U with four and 4U with eight Pascal P100 SXM2 GPUs or 4U with ten PCI-e GPU systems, hot-swap U.2 NVMe, upcoming fabric technologies like Red Rock Canyon and PCI-E switches, as well as new architecture designs like our new high-density BigTwin system design,” said Charles Liang, President and CEO of Supermicro. “Our total end-to-end solutions offer exactly the best deployment options for supercomputing clusters, delivering maximum performance per watt, per square foot, and per dollar.”