JakeATX/llamAmpere

llamAmpere

✅ Recently maintained

llama.cpp fork for significantly improved performance on Ampere generation Nvidia cards (especially RTX 3090 / 3090 Ti): example: 110+ tok/s over a 100K-token generation at temperature 1 for qwen3.8, supporting full 262K ctx. Supports low memory (12gb) configs for 200k+ ctx

214Stars
30Forks
+267-day stars

C++ · MIT · Last push 2026-11-03

Official repository & installation ↗
ampereconsumer-gpucudaggmlggufkv-cache-quantizationllama-cppllmllm-inferencelocal-llm

GitValue · Value evidence

Unscored · Evidence coverage 0%

Provisional rubric; documentation and runtime testing are separate. Missing evidence does not imply low value.

Star history

2026-09-072026-11-03
View historical data
DateStars
2026-10-05142
2026-10-06146
2026-10-07146
2026-10-08147
2026-10-09155
2026-10-10164
2026-10-11166
2026-10-12168
2026-10-13170
2026-10-14173
2026-10-15175
2026-10-16176
2026-10-17178
2026-10-18179
2026-10-19180
2026-10-20182
2026-10-21183
2026-10-22184
2026-10-23185
2026-10-24186
2026-10-25187
2026-10-26188
2026-10-27189
2026-10-28196
2026-10-29202
2026-10-30203
2026-10-31203
2026-11-01207
2026-11-02214
2026-11-03214

Get the project

git clone https://github.com/JakeATX/llamAmpere.git

Cloning downloads the source. For dependencies and installation, follow the project README.

Related projects & alternative candidates

Related domains do not imply feature parity. Review official documentation and licenses.

nanobot

Lightweight swarm intelligence experimental platform for micro-agents.

★ 49,776vllm

High-throughput, memory-efficient LLM inference and serving engine that lets developers deploy models faster without building serving infrastructure.

★ 96,180llmfit

Rust CLI tool that scans your local hardware specs and recommends which LLMs and quantization formats will run smoothly.

★ 38,587