Overview
Enterprise 8u rackmount server featuring 8x NVIDIA B300 GPUs with HGX platform. Designed for next-generation AI training with Blackwell architecture delivering unprecedented FP4/FP8 performance for large language model workloads.
Key Features
- 8x NVIDIA B300 GPUs on HGX baseboard
- NVLink 5.0 interconnect for 1.8TB/s GPU-to-GPU bandwidth
- Blackwell architecture with FP4/FP8 transformer engine
- Up to 192GB HBM3e memory per GPU
- Enterprise-grade redundant power and cooling
Ideal For
Large enterprises and hyperscalers requiring maximum AI training performance for frontier model development.
Feature
Dual IntelÂŽ XeonÂŽ 6700 series processors with P-cores, up to 350W TDP
Feature
8x CX-8 Integrated 800G NICs supporting IB XDR or ETH 2x400GbE per NIC
Feature
32x DIMM slots supporting up to 8TB of memory
Feature
1x USB 3.0, 1x USB 2.0, 1x Mini-DP, 1x BMC/IPMI GbE LAN, 2x 10GbE NIC (X710)
$760,000.00
Prices may vary. Verify on vendor site.
Quick Specs
- Feature
- 1x USB 3.0, 1x USB 2.0, 1x Mini-DP, 1x BMC/IPMI GbE LAN, 2x 10GbE NIC (X710)
Tags
llm-traininggenerative-aihpc