NVIDIA · Ampere
Switzerland · 48 vCPU · 225 GB RAM · Storage billed separately
Stock and price checked 14 min ago.
You start it yourself once there is stock.
This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.
gpua30.huge
Switzerland · 24 vCPU · 120 GB RAM · Storage billed separately
Stock and price checked 14 min ago.
You start it yourself once there is stock.
This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.
gpua30.large
Switzerland · 16 vCPU · 90 GB RAM · Storage billed separately
Stock and price checked 14 min ago.
You start it yourself once there is stock.
This GPU supports NVLink 3, but the provider doesn't say whether this server has it connected. Ask before relying on it.
gpua30.medium
Switzerland · 12 vCPU · 56 GB RAM · Storage billed separately
Stock and price checked 14 min ago.
You start it yourself once there is stock.
gpua30.small
Excluding VAT.
Each provider counts once. Full price and spot are separate.
Full price. Offers we can't say are really sold there are left out until they've been in stock once.
No provider reports a number for this GPU.
In stock, yes or no, for every provider selling this GPU. A filled band means in stock: green at full price, orange as spot.
Can be linked in pairs with an NVLink 3 bridge at up to 200 GB/s, if the provider has fitted one. Most providers don't say, so ask before relying on it. Software still has to split the work, so the memory doesn't act as one big pool.
NVLink bridge connects up to two GPUs; datasheet gives no CUDA core count. memoryDataRateGbps is derived (933 GB/s x 8 / 3072 bits, matching the 1215 MHz memory clock x 2). interconnectLinks is derived from 32 NVLink Rx+Tx lanes at 8 lanes per link (A100 uses 96 lanes for 12 links); per-link 50 GB/s = 200 GB/s / 4.