GPU Server
Supermicro’s powerful GPU servers push most compute-intensive workloads to new performance levels, delivering top quality through high optimization. They provide massive parallel processing capability and exceptional networking flexibility, supporting up to 8/4 GPUs or coprocessors, PCI-E 3.0 expansion slots, 10GBase-T Ethernet, and QDR/FDR InfiniBand ports. These systems are built with the latest and most advanced technologies.
Supermicro GPU servers support: up to 6TB ECC 3DS LRDIMM or RDIMM DDR4-2933MHz memory (in 4 DIMM slots), Intel® Optane™ DC Persistent Memory, 1U / 2U / 4U form factors, 4 SATA 3.0 (6Gbps), 4/6 GPU or coprocessor card ports, Intel® C622 controller, up to 2 10GBase-T ports, redundant Platinum-level (95%) power supplies, dedicated Redfish LAN, and dual or single 2nd Gen Intel® Xeon® Scalable processors with up to 28 cores and 255W TDP.
Supermicro’s powerful GPU servers are available in 1U, 2U, 4U, and tower form factors, optimized for HPC, AI / deep learning, oil and gas simulation, computational finance, science and engineering, and media and entertainment.
Key Features
-
Form Factor
1U, 4U rackmount
-
Memory
Up to 24 DIMM slots, up to 6TB DDR4 memory, Optane™ persistent memory
-
Power
Redundant Titanium (96%) and Platinum power efficiency options
-
Management
Open industry-standard IPMI, Redfish APIs, rack-scale management
-
Processors
Dual Intel® Xeon® Scalable processors up to 205W TDP, or dual AMD EPYC™ processors
-
Drives
Supports up to 16x U.2 NVMe drives, with 2.5" and 3.5" SAS3 / SATA3 options
-
Input/Output
Up to 4 GPUs in 1U, up to 20 GPUs in 4U
