C++ · GPL-3.0
1115714829/llama.cpp-v100
llama.cpp optimized for NVIDIA Tesla V100 (SM70). Qwen3.8-Flash-Next (MoE) with MTP on 1/2/3/6x V100-16GB at 262K context (6 GPUs: prefill 3402 tok/s at 200K; 2 GPUs on one socket: 2152 tok/s). Qwen3.8-27B with DFlash2 on 1-6 GPUs, up to 512K context. Multimodal, thinking, concurrent serving, built-in web UI.
★ 71
Stars
14 forks · 8 open issues
+1.4/day
Growth · measured
Stars counted between two of our own snapshots roughly a week apart. The repository page gives the exact span.
27.6
Momentum / 100
see the breakdown below
48d ago
Created
last push 5d ago
Star history
How this score was built
| Component | Weight | Earned | Why |
|---|---|---|---|
| Velocity | 55 | 7.4 | +1.4/day, log-scaled against a 750/day ceiling |
| Acceleration | 15 | 0.0 | last week ran 0.90× the monthly pace |
| Freshness | 15 | 14.8 | last push 5d ago; decays to zero at 180 days |
| Scale | 15 | 5.4 | 71 stars, log-scaled against 150k — stops a 3-star repo topping the board |
| Momentum | 100 | 27.6 | sum of the above |
Detail
| Topics | cudaggufllama-cppllm-inferencelocal-llmlong-contextmixture-of-expertsmoemtpmulti-gpumultimodalqwenqwen3sm70speculative-decodingtensor-parallelismtesla-v100v100volta |
| Matched on | LLM & Inference (llm-inference) |
| Short window | +10 stars over 7 days = 1.4/day |
| Long window | +27 stars over 17 days = 1.6/day |
| Lifetime average | 1.57 stars/day over 45 days |
| First seen here | 2026-10-19 |
| Last read | 3d ago |