rtx-3090 repositories

Trending repositories tagged rtx-3090, ranked by trending score. Updated daily.

  1. 1
    JakeATX/llamAmpere

    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. 120+ TPS, 3300+ prefill for QFN 3.8.

    GitHub · C++ · 403 stars · 50 forks · +87 stars in 24 hours · Trending score 3.35

  2. 2
    NODIX-TECH/DLSS-5-MANAGER

    THE FIRST DLSS 5 MANAGER

    GitHub · F# · 1,636 stars · 78 forks · +18 stars in 24 hours · Trending score 2.45

  3. 3
    bpx03/QEngine

    Fast open-source (MIT) LLM inference engine in C++/CUDA for RTX 3090-class GPUs: Qwen3.8-Flash-Next (~1,000 tok/s with 8 agents on 3x RTX 3090) and Qwen3.8-27B, OpenAI/Anthropic-compatible API.

    GitHub · Cuda · 75 stars · 8 forks · +4 stars in 24 hours · Trending score 1.75

  4. 4
    BrightRaider/universal-gpu-unlock

    Universal one-click UEFI pre-boot rate-limiter unlock for NVIDIA GeForce RTX 40 & 30 Series (Ada Lovelace & Ampere). Uncaps full tensor performance across Windows, PyTorch, CUDA, and WSL2.

    GitHub · C · 22 stars · 6 forks · +3 stars in 24 hours · Trending score 1.29

  5. 5
    tuan3w/qwen-warp

    Fast Qwen3.8-27B for one RTX 3090. 200k context, 100+ tok/s. Custom CUDA decode engine, OpenAI-compatible API.

    GitHub · Python · 49 stars · 6 forks · +3 stars in 24 hours · Trending score 1.16

  6. 6
    syv-ai/HyperQwen

    Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when the answer quotes the prompt), ~1,035 tok/s at 64 concurrent, 150k-262k context. vLLM patches, requant pipeline, benchmarks.

    GitHub · Python · 2,696 stars · 340 forks · 0 stars in 24 hours · Trending score 0.61

  7. 7
    tabilzad/gpu-unlock-unraid

    Unraid plugin: unlock the full FP16/BF16 rate of RTX 30 series cards at boot with Zagitta/gpu-unlock

    GitHub · Shell · 11 stars · 1 forks · 0 stars in 24 hours · Trending score 0.47

  8. 8
    alphastorm/omp-ninfer

    Durable local inference for Oh My Pi: NInfer + Qwen3.8 27B on one RTX 3090/4090/5090, with restart-resumable OpenAI Responses state.

    GitHub · Python · 50 stars · 5 forks · 0 stars in 24 hours · Trending score 0.28

  9. 9
    alesha-pro/qwen38-flashnext-strata-4x3090

    Qwen3.8-Flash-Next on 2x RTX 3090: 150 tok/s on code, 262K context. Strata 0.2.10 + unsloth UD-IQ4_XS + DFlash2. Full recipe, refusal vectors, session patch, raw benchmark traces.

    GitHub · Python · 10 stars · 1 forks · 0 stars in 24 hours · Trending score 0.14

  10. 10
    mikecovlee/qwen3.8-27b-24gb-recipe

    Run Qwen3.8-27B locally on any 24GB GPU (RTX 3090/4090/A5000) with llama.cpp: install, scripts, MTP tuning, benchmarks

    GitHub · PowerShell · 13 stars · 2 forks · 0 stars in 24 hours · Trending score 0.04

  11. 11
    devnen/qwen3.6-windows-server

    One-click Qwen3.6-27B inference on Windows. 158 tok/s on RTX 5090, 72 tok/s on RTX 3090. Native, no WSL, no Docker, no telemetry.

    GitHub · Python · 196 stars · 22 forks · 0 stars in 24 hours · Trending score 0.00

  12. 12
    alesha-pro/tools

    Tools, ComfyUI workflows and benchmark configs from a 4x RTX 3090 local-inference rig

    GitHub · HTML · 687 stars · 77 forks · 0 stars in 24 hours · Trending score 0.00

  13. 13
    halt95/merlin-engine

    vLLM 0.32 fork for consumer GPUs: 3 x 262K sessions on 4x RTX 3090, Qwen 3.8 Flash-Next at 197 tok/s, up to 17 on bigger cards. Seven checkpoints incl. Qwen3.8-27B, Nemotron 3.5 Super VL (1M) and Ornith 1.5. One command: merlin serve.

    GitHub · Python · 42 stars · 7 forks · 0 stars in 24 hours · Trending score 0.00