3 ms·
If people are interested, I could split up each of our MI300x into 4 and then charge $0.50 (1/4th our current rate). You'd get 48GB of vram instead of 32GB and
by latchkey 1y ago
If people are interested, I could split up each of our MI300x into 4 and then charge $0.50 (1/4th our current rate). You'd get 48GB of vram instead of 32GB and it would be HBM3 instead of GDDR7 (5.3TB/s vs 1.7TB/s).
The only catch is that I'd need to get 32 people who want VMs like this since I would have to do it for the entire box of compute.
Wan2.2 runs just fine on AMD.
- ricardobeat 1y agoVultr does this: https://www.vultr.com/pricing/#cloud-gpu https://www.vultr.com/pricing/#cloud-gpu Though only a shared A40/A100 are in that price range.
- latchkey 1y agoIt really is an unfortunate thing that pricing is so opaque and non-transparent in this industry. You look at one price and that's it. The reality is more complex. Vultr is a box of 8 minimum and not on-demand and they don't offer VMs. On the other hand, I offer the bare minimum (1 GPU for 1 minute) (or 2, 4, 8x), on-demand, no-contract, and an API to automate it all. We also have 100G unlimited bandwidth and free IPv4. Oh and our 8x box specs are generally better... 122TB of enterprise NVMe.
- ricardobeat 1y agoWhat's your company? Consider adding it to your profile.
- latchkey 1y agodm'd