Supermicro Showcases New GPU SuperServer at GTC 2016

Print Friendly, PDF & Email

SYS-1028GQ-TRToday SuperMicro rolled out its next generation GPU server architecture alongside new GPU optimized SuperServer, SuperBlade, and SuperWorkstation solutions this week at the NVIDIA GPU Technology Conference (GTC). Supermicro exhibits at GTC include 1U GPU SuperServers supporting up to 4x NVIDIA Tesla GPU accelerators, 2U 2x GPU TwinPro, 4U 12x GPU FatTwin, 4U 8x GPU, 4U/Tower 4+1 GPU SuperServers and 7U SuperBlade 20x GPU system solutions.

Supermicro’s new SuperServer and SuperBlade platforms optimized for next generation GPU technology offer the most flexible architecture for the future of extreme parallel computing,” said Charles Liang, President and CEO of Supermicro. “Our proprietary cooling architecture supports up to four GPUs in 1U, maximizing compute density, performance and reliability under peak thermal operating conditions. Supermicro’s innovation and revolutionary solutions deliver the greatest range of options for next generation supercomputers, enabling new breakthroughs in the latest deep learning and artificial intelligence applications.”

From HPC to Deep Learning and Big Data Analytics, denser, more powerful GPU solutions have become a necessity in order to service the next generation of GPU-accelerated applications. At GTC, Supermicro will demonstrate how these applications have progressed, and how its GPU solutions are influencing this evolution. Supermicro’s unique Green Computing architecture features redundant high-efficiency power supplies and airflow optimized cooling architecture to maximize thermal operating conditions and increase energy efficiency and reliability. With Supermicro GPU-accelerated computing solutions, engineering, scientific and research fields can scale out compute clusters that accelerate their most demanding workloads and achieve fastest time to results with maximum performance per watt, per square foot, per dollar.

Supermicro GTC Exhibits include:

  • 1U 4x GPU SuperServer – supports up to 4x NVIDIA Tesla GPU accelerators with innovative non GPU-preheat architecture, dual Intel Xeon processor E5-2600 v4, up to 2TB DDR4 2400MHz ECC DIMMs, 2 hot-swap and 2 fixed 2.5” drive bays, and redundant 2000W Titanium Level high-efficiency (96%) power supplies
  • 1U 3x GPU SuperServer – supports up to 3x NVIDIA Tesla GPU accelerators, dual Intel Xeon processor E5-2600 v4, up to 2TB ECC, DDR4 2400MHz in 16x DIMM slots, 4 hot-swap 2.5″ SATA3 drive bays, 4 PCI-E 3.0 x16 slots plus 1 PCI-E 3.0 x8 LP slot, dual 10GBase-T ports, and redundant 1600W Platinum Level (94%+) power supplies
  • 2U TwinPro 2x GPU SuperServer – Two hot-pluggable systems (nodes), each node supports dual Intel Xeon processor E5-2600 v4 family; QPI up to 9.6GT/s, up to 2TB ECC LRDIMM, 512GB ECC RDIMM, up to 2400MHz; 16x DIMM slots, 1x PCI-E 3.0 x16, 1x PCI-E 3.0 x8 slot, and 1x PCI-E 3.0 x16 for single NVIDIA Tesla GPU accelerator support (w/ GPU kit), single port IB (FDR, 56Gbps) w/ QSFP connector, dual port GbE LAN, 12x 2.5″ Hot-swap SAS (8) / SATA (4) HDD Bays, LSI 3108 SAS3 controller (8 ports); RAID 0, 1, 5, 6, 10, 50, 60, mSATA (full size) support, 1280W Redundant Platinum Level (94%) digital power supplies
  • 4U FatTwin 12x GPU SuperServer – 4x Hot-plug System Nodes in 4U w/Front I/O, each node supports dual Intel Xeon processor E5-2600 v4 family, up to 2TB ECC DDR4, up to 2400MHz in 16x DIMM slots, 3x PCI-E 3.0 x16 slots (support 3x double-width NVIDIA Tesla GPU accelerators), 2x PCI-E 3.0 x8 slots, Front I/O ports: 2 10GBase-T, 2x USB 3.0 and 1 VGA connector, 2x 3.5″ Hot-swap SAS3/SATA3; SAS3 (12 Gbps) via LSI 3008, Redundant 2000W Titanium Level High Efficiency (96%) digital power supplies
  • 4U 8x GPU SuperServer – supports dual Intel Xeon processor E5-2600 v4 family; up to 3TB ECC DDR4, up to 2400MHz ; 24x DIMM slots, 8x PCI-E 3.0 x16 (double-width) slots supporting NVIDIA Tesla GPU accelerators, 2x PCI-E 3.0 x8 (in x16) slots, 1x PCI-E 2.0 x4 (in x16) slot, dual 10GBase-T LAN with Intel® X540, 24x 2.5″ Hot-swap drive bays, 1600W Redundant (2+2) Platinum Level (94%+) power supplies
  • 4U/Tower 4+1 GPU SuperServer – supports 4+1 NVIDIA GPU accelerators, NVIDIA SLI technology, dual Intel Xeon processor E5-2600 v4, up to 2TB ECC DDR4 2400MHz in 16x DIMM slots, 8x 3.5″ hot-swap, 3x fixed 5.25″ and 1x fixed 3.5″ drive bays, 4x heavy duty fans, 2x external exhaust fans, and 2x active heatsink with optimal fan speed control, redundant 2000W Titanium Level high efficiency (96%) digital power supply, passive GPU kit (MCP-320-74701-0N-KIT) and optional Thunderbolt 2.0 AOC
  • 7U SuperBlade – features 120 GPU + 120 CPU per 42U Rack! Supports up to 2x NVIDIA GPU accelerators, dual Intel Xeon processor E5-2600 v4, up to 1TB ECC DDR4 2400MHz in 8x DIMM slots, 2x SATA DOM or 1x SSD; or up to 8x 2.5” SATA3 HDDs + 1x 2.5” SSD, onboard 10GbE (-X), single FDR IB port (-F), dual FDR IB port (-F2)

Visit Supermicro at the GPU Technology Conference April 4-7, Booth #818.

Sign up for our insideHPC Newsletter