Overview
2u rackmount server and 9x H100 80GB GPUs. Designed for large-scale AI training and multi-GPU deep learning workloads.
Key Features
- 9x H100 with 80GB VRAM
- 2u rackmount form factor
- Enterprise-ready configuration
Ideal For
enterprise teams and research institutions running large-scale AI training, multi-GPU workloads, and enterprise inference deployments.
System
Supermicro - 821GE-TNHR
CPU
SPR 8468V 2P 48C 2.4G 330W (2.1/300,1.8/270) 97.5MB
Memory
64GB 4800MHz DDR5 ECC RDIMM
SSD
SSD 2.5ā³ SATA 480GB >1DWPD TLC E, 7mm
NVMe SSD
SSD 2.5ā³ NVMe PCle4 7.6TB 1DWPD TLC D, 15mm
GPU
NVIDIA DELTA-NEXT HGX GPU Baseboard, 8 H100 80GB SXM5
Network Card
Nvidia 900-9X766-003N-SQ0 PCle 1-port IB 400GE OSFP
AOC
Std LP 2-port 10G RJ45, Intel X550
TPM
Supermicro -SPI TPM2.0 using SLB9670
ŠŠB
(Supermicro Management Software Suite Node License)
Kit
SOCKET E LGA4677,E1A XCC CARRIER,DG 1.0,RoHS
Warranty
3 Years Return to Base Warranty
$289,000.00
Prices may vary. Verify on vendor site.
Quick Specs
- System
- Supermicro - 821GE-TNHR
- CPU
- SPR 8468V 2P 48C 2.4G 330W (2.1/300,1.8/270) 97.5MB
- Memory
- 64GB 4800MHz DDR5 ECC RDIMM
- SSD
- SSD 2.5ā³ SATA 480GB >1DWPD TLC E, 7mm
- NVMe SSD
- SSD 2.5ā³ NVMe PCle4 7.6TB 1DWPD TLC D, 15mm
- GPU
- NVIDIA DELTA-NEXT HGX GPU Baseboard, 8 H100 80GB SXM5
Tags
llm-traininggenerative-aihpcsimulation