
TARGET: Enterprise Level – GPUDirect RDMA Applications AI Neural Network workloads
G4760
4U 10 GPU Server
- Single Root Architecture
- Supports up to 10 Double or 20 Single Wide GPU’s
- Nvidia Tesla V100, P100 & P4 selection
- Intel® Xeon® Scalable Processors
- Memory expandable to 3 TB
- Dual 10GbE LAN support – options to multiple 100 GbE
- Fourteen 2.5” Hot Swap SSD/HDD Bays
- Open BMC w/Remote Management – Redfish support
The G4760 is a dense 10GPU single root platform that is highly optimized for Deep Learning and Artificial Intelligence workloads. Each GPU is connected to the same CPU using PCIe switches to handle the connectivity. This GPUDirect RDMA topology is being used in many of today’s deep learning applications that are being developed on single root systems to realize potential advantages in bi-directional bandwidth and latency. The G4760 has up to 14 2.5” hot-swap SSD/HDD drive bays with support for 10 + 4 NVMe, up to 3TB memory, (3+1) 4,800W RPSU, and a solid mechanical structure. This GPU platform provides the power and flexibility for high performance AI applications.
Product Specifications
Chassis Dimensions | 4U Rackmount : 30.31" x 17.24" x 6.93" |
Processor | Dual Intel® Xeon® Scalable up to 205W |
Chipset | Intel® C621 Chipset |
Drive Bays - External | 14x 2.5" Hot-Swap SDD/ HDD Bays |
Drive Bays - Internal | None |
Backplane | SAS 12Gb/s, SATA 6.0Gb/s |
Memory | Up to 768GB RDIMM/ 1,536GB LRDIMM/ 3,072GB RDIMM |
Expansion Slots | (11) PCI-E Gen3 x16 slots |
Riser Card | M2217-L16-IO |
OCP | None |
Network Controllers | Intel® X540-BT2 2x 10GbE, 1x 1GbE dedicated for IPMI |
Graphics | Aspeed AST2500, up to 1920x1200, D-sub 15-pin connector |
Rear I/O | 2x RJ-45 10GbE, 1x GbE for IPMI, 2x USB3.0 at rear, 1x D-SUB VGA, 1x DB-9 |
System Management | AST2500 BMC Chip, IPMI 2.0 compliant |
Power Supply | RPSU 2300W hot swap, 80-PLUS Platinum |
Environmental Specs | Operating: 10° C ~ 35° C (50° F~ 95° F) |
Gross Weight | 45 kg (99.5 lbs) |
OS Support | W2k12 R2 Datacenter x64 W2k12 Enterprise R2 x64 W2k16 RHEL 6.9 x64, 7.3 x64, & 7.4 x64 Ubuntu Server 16.04 x64 |