posted by Hypercompute 20 days ago

$2.00 / gpu / hour - 64 H100s available from 05/31/2024 to 05/31/2025 (US West)

Unverified
Interconnect: Ethernet
Cores per node: 32
RAM per node: 1500 GB
NVME storage per node: 12000 GB
Location: US West
Cloud provider: 10GbE
Cluster interface: K8s
Minimium bookable GPUs: 8
Minimium bookable duration: 12 weeks
Additional details
Available end of May.
H100 4U Rackmount 10x GPU Capable System
2x AMD EPYC Genoa 9124 - 16-Core, 32-Threads, 3.0GHz Base, 64MB Cache - Default TDP: 200w cTDP: 200-240w 64GB DDR5 24x 4800M/Ts ECC REG Memory Module ( -- 1.5TB TOTAL RAM -- )
3x 3.84TB U.2/U.3 2.5in NVMe PCIe 4.0 Enterprise SSD
--- OS: 1x 3.84TB NVME SSD
--- SCRATCH: 1x 3.84TB NVME SSD ---
DATA: 1x 3.84TB NVME SSD
8x NVIDIA H100 Hopper NVL 94GB Graphics Card - PCIe 5.0 - 350-400W
3x MCX653105A-HDAT - NVIDIA ConnectX-6 VPI Adapter Card HDR IB (200Gb/s) and 200GbE Single-Port QSFP56 PCIe4.0 x16 Tall Bracket ROHS R6
Ubuntu 22.04 LTS
contact seller

seller email domain is hypercompute.ai