posted by Gradient 5 months ago
$2.40 / gpu / hour - 64 H200s available from 12/31/2024 to 12/31/2026 (San Jose, CA)
Unverified
Interconnect: Infiniband 3.2 Tb
Cores per node: 112
RAM per node: 2000 GB
NVME storage per node: 38 GB
Location: San Jose, CA
Cloud provider: N/A
Cluster interface: N/A
Minimium bookable GPUs: 8
Minimium bookable duration: 4 weeks
Additional details
DGX H200s available for rental end of December. We are offering a 5-15% discounts for annual commitments before December 31.
--- 8x NVIDIA H200 SXM GPUs with 141GB for 1.128TB of Total GPU Memory
--- 18x NVIDIA NVLink connections per GPU, 900 gigabytes per second of bidirectional GPU-to-GPU bandwidth
--- 4x NVIDIA NVSwitches w/ NVLink Connecting all 8x GPU
--- Dual Intel® Xeon® Platinum 8480C Processors For 112 Cores total, 2.00 GHz (Base), 3.80 GHz (Max Boost)
--- 2TB Total RAM
--- 8x 3.84TB NVMe SSD for 30TB of Storage
--- 2x 1.92TB NVMe SSD for OS
--- 4x 800G OSFP External ports serving 8x NVIDIA ConnectX-7 VPI 400G Connections Internally
--- 2x dual-port QSFP112 NVIDIA ConnectX-7 VPI
--- 10Gb/s onboard NIC with RJ45
--- 100Gb/s Ethernet NIC
--- 1x Host baseboard management controller (BMC) with RJ45
contact seller
seller email domain is gradient.ai