Can I run Ripple-65B-Preview-GGUF on an NVIDIA H200?
Yes
Its Q8_0 file (68.8 GB) fits in the 127 GB an NVIDIA H200 can use, with room for ≈ 861K tokens of context.
Full analysis of Ripple-65B-Preview-GGUF
Everything that runs on NVIDIA H200
Get the Q8_0 file on Hugging Face
llama-server -hf AtomicChat/Ripple-65B-Preview-GGUF:Q8_0Files and context on NVIDIA H200
Each file of Ripple-65B-Preview-GGUF (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.
| File | Size | On NVIDIA H200 | Context (16-bit) | Context (8-bit) |
|---|---|---|---|---|
| BF16 | 130 GB | Doesn't fit | — | — |
| Q8_0 | 68.8 GB | Fits | 861K | 1.7M |
| Q6_K | 57.7 GB | Fits | 1.0M | 2M |
| Q5_K_M | 47.9 GB | Fits | 1.1M | 2M |
| Q4_K_M | 40.4 GB | Fits | 1.3M | 2M |
| Ripple-65B-Preview-Q4_K_M | 38.8 GB | Fits | 1.3M | 2M |
| IQ4_XS | 38.7 GB | Fits | 1.3M | 2M |
| IQ3_S | 32.4 GB | Fits | 1.4M | 2M |
| IQ3_XXS | 28.0 GB | Fits | 1.4M | 2M |
| IQ2_S | 25.3 GB | Fits | 1.5M | 2M |
| IQ2_XS | 22.1 GB | Fits | 1.5M | 2M |
| IQ2_XXS | 20.0 GB | Fits | 1.6M | 2M |
| IQ1_M | 18.5 GB | Fits | 1.6M | 2M |