H100 SXM5 HGX

NVIDIA HGX H100 combines H100 Tensor Core GPUs with high-speed interconnects to form the world’s most powerful accelerated scale-up server platform for AI and HPC. With up to eight H100 GPUs, HGX H100 has up to 640 gigabytes (GB) of GPU memory and 24 terabytes per second (TB/s) of aggregate memory bandwidth for unprecedented acceleration. HGX H100 includes advanced networking options—at speeds up to 400 gigabits per second (Gb/s)—utilizing NVIDIA Quantum-2 InfiniBand and Spectrum™-X Ethernet for the highest AI performance. HGX H100 also includes NVIDIA® BlueField®-3 data processing units (DPUs) to enable cloud networking, composable storage, zero-trust security, and GPU compute elasticity in hyperscale AI clouds.

Our optimized HGX configuration

  • CPU: Dual Intel 8480+ Processors

  • GPU: 8x H100 80GB (640GB) SXM5

  • Memory: 32x 64GB DDR5 4800Mhz ECC (2048GB total)

  • Boot Drive: 960GB M.2 NVMe

  • SSD Storage: 30TB NVMe (4x 7.68TB or 8x 3.84TB)

  • RDMA Network Card: 8x ConnectX-7 400G InfiniBand SP

  • Storage Network Card: Dual Port 100G

  • TPM: TPM-SPI 2.0