Question 1

Can the NVIDIA H100 80GB PCIe run mistral 7B?

Accepted Answer

mistral 7B needs about 4.6GB working set (weights 4.1GB + KV cache 0.5GB at 4k context). The NVIDIA H100 80GB PCIe offers ~80GB usable — verdict: full. 4.6 GB needed of 80 GB usable — headroom for context

Question 2

How fast is mistral 7B on the NVIDIA H100 80GB PCIe?

Accepted Answer

~230–380 tok/s on the NVIDIA H100 80GB PCIe. computed roofline: 2039 GB/s × 0.46–0.76 efficiency window / 4.1GB (Q4_K_M) — bands, never points.

Question 3

What should I run mistral 7B on instead?

Accepted Answer

Tesla M60 (16GB, used) runs it fully ($35).

Can the NVIDIA H100 80GB PCIe run mistral 7B?

Run mistral 7B on instead