Question 1

Can the NVIDIA H100 80GB PCIe run DEEPSEEK-R1 distill 70B?

Accepted Answer

DEEPSEEK-R1 distill 70B needs about 44.25GB working set (weights 43GB + KV cache 1.25GB at 4k context). The NVIDIA H100 80GB PCIe offers ~80GB usable — verdict: full. 44.2 GB needed of 80 GB usable — headroom for context

Question 2

How fast is DEEPSEEK-R1 distill 70B on the NVIDIA H100 80GB PCIe?

Accepted Answer

~22–36 tok/s on the NVIDIA H100 80GB PCIe. computed roofline: 2039 GB/s × 0.46–0.76 efficiency window / 43.0GB (Q4_K_M) — bands, never points.

Question 3

What should I run DEEPSEEK-R1 distill 70B on instead?

Accepted Answer

NIMO AI Mini PC (128GB) runs it fully ($1999).

Can the NVIDIA H100 80GB PCIe run DEEPSEEK-R1 distill 70B?

Run DEEPSEEK-R1 distill 70B on instead