Find more information about the provisioning limitations, which are primarily determined by the physical hardware constraints.
These constraints are relevant to GPU instances that match the specific GPU model. As the number of GPUs increases, vCPU per GPU, RAM per GPU, and storage per GPU also increase proportionally.
GPU
|
Max GPUs per Instance
|
vCPUs per GPU
|
RAM per GPU
|
Storage per GPU
|
Public Network
|
---|---|---|---|---|---|
H100 80GB NVLINK | 8 | 32 | 180 | 825GB NVMe | 100Gbps |
H100 80GB PCIE | 8 | 28 | 180 | 850GB NVMe | 100Gbps |
A100 80GB NVLINK | 8 | 32 | 240 | 825GB NVMe | 100Gbps |
A100 80GB PCIE | 8 | 28 | 120 | 850GB NVMe | 25Gbps |
L40 | 8 | 28 | 58 | 825GB NVMe | 25Gbps |
RTX A6000 | 8 | 16 | 60 | 425GB NVMe | 25Gbps |
RTX A5000 | 8 | 8 | 30 | 200GB NVMe | 10Gbps |
RTX A4000 | 8 | 6 | 24 | 160GB NVMe | 10Gbps |