Super Micro Computer, Inc.†(SMCI), a global leader in enterprise computing, storage, networking solutions, and green computing technology, is demonstrating the industry’s broadest selection of new high-performance computing (HPC) systems at SuperComputing 2018 (SC18).
“As a hardware solution company, we are investing heavily in our Resource-Saving server, GPU and storage solutions, including the development of 10-year lifecycle chassis, power supplies, fans, and other subsystems, to help end-customers save both energy cost and hardware acquisition costs while reducing IT waste,” said†Charles Liang, President and CEO of Supermicro. “Resource Saving is measured by TCE (Total Cost to the Environment) which is the combination of delivering superior TCO for datacenter investments while at the same time minimizing the environmental impacts of these data centers. Also, continuing Supermicro’s technology innovation and time-to-market leadership, our Petascale 1U NVMe solutions are shipping in volume to provide customers with significant competitive advantage today.”
The new Petascale line of all-flash NVMe (Non-Volatile Memory Express) 1U storage servers feature the most power-efficient, next-generation flash technology with the highest storage density and best IOPS performance.† With systems supporting U.2, Intel Ruler, Samsung NF1, and EDSFF form factor SSDs, Supermicro offers unprecedented flexibility and choice for high-capacity networked storage applications that require the best latency performance and provide a real time-to-value advantage for users with data-intensive HPC workloads.
HPC applications are continuing to grow in complexity as they unlock new scientific insights. Supermicro’s new NVIDIA†HGX-2 based SuperServer, 9029GP-TNVRT, supports 16 NVIDIA Tesla†V100 Tensor Core†32GB†GPUs connected via NVIDIA NVLink†and NVSwitch†to leverage over 80,000 CUDA cores and delivers unmatched performance accelerating AI and HPC on-premise and in the cloud. This new system can deliver up to 2 PetaFLOPS of performance and occupies just ten units of rack space.†
To help simplify the deployment of AI and HPC applications in the data center, SuperServer 4029GP-TVRT is NGC-Ready. Customers can now run GPU-accelerated software from the NGC container registry, including its expanded HPC and AI software library with new machine learning and analytics containers, with confidence on the 4029GP-TVRT, which is a 4U system with eight Tesla V100 GPUs with NVIDIA NVLink.
Designed to handle the most demanding inference workloads, Supermicro’s new SuperServer 6049GP-TRT provides the superior performance required for modern AI.† This 4U system achieves maximum GPU density and performance with support for up to 20 NVIDIA Tesla T4 GPUs with Turing Tensor Core technology, three terabytes of memory, and 24 hot-swappable 3.5-inch drives. The system also features four 2000-watt Titanium level efficiency (2+2) redundant power supplies to deliver optimal power efficiency, uptime, and serviceability.
Delivering maximum performance and efficiency in a 2U 4-node design, the†Supermicro BigTwin system†supports the full range of Intel Xeon Scalable processors, 24 DIMMs, up to six hot-swappable all-flash NVMe or hybrid NVMe/SAS3 drive bays, and up to three PCI-E 3.0 slots including support for a flexible SIOM module enabling 100/40/25/10/1G networking options per node. For reliability and efficiency, redundant 2600W/2200W Titanium Level (96%+) digital power supplies come standard. Supermicro also offers a line of BigTwin systems that support the full range of AMD EPYC 7000-series processors including an all-flash NVMe model.
Supermicro is showcasing a breadth of platforms at SC18 that feature Resource-Saving technology and address a wide range of workloads. The new disaggregated SuperBlade unlocks the interdependence between the major server subsystems enabling the independent upgrade of CPU, Memory, I/O, Enclosure, Storage, Power, and Cooling.† Now each component can be refreshed at the optimal time to maximize generational improvements in performance and efficiency. SuperBlade delivers maximum performance up to 495 TFLOPS per rack, high density with twenty 2-socket or ten 4-socket Intel Xeon Scalable processor-based blade servers per 8U enclosure, fast interconnects with integrated 100G EDR InfiniBand, 100G Omni-Path or 25G switches, and support for up to 14 NVMe devices per node. For optimal reliability and power-efficiency, SuperBlade supports redundant 2200-watt Titanium Level (96%+) digital power supplies, 2000-watt DC power supplies, and 1200-watt Battery Backup Power (BBP) modules.
