Popular repositories Loading

1. beellama-cmp90hx beellama-cmp90hx Public

   beellama.cpp fork with CMP 90HX (GA102, sm_86) optimizations: DP4A→IMAD, FP32→HFMA2 — +58-87% decode speed

   C++ 10 3

2. inferrank inferrank Public

   Forked from noonghunna/club-3090

   Budget GPU LLM inference leaderboard — RTX 3060, CMP 90HX, mining cards, AMD. Real benchmarks, 150 scenarios, multiple quality packs.

   Python 6

3. cmp90hx-unlock cmp90hx-unlock Public

   Forked from WildFlash1st/cmp90hx-unlock

   Unlocking NVIDIA CMP 90HX (GA102): kernel 6.12 port + Falcon ROP exploit research. Contributors wanted!

   Python 3

4. CMP-90HX-Compute-PCIE2.0-unlock CMP-90HX-Compute-PCIE2.0-unlock Public

   Forked from NatsumeAi/CMP-90HX-Compute-PCIE2.0-unlock

   CMP 90HX compute + Gen2 unlock (VBIOS .07) — fleet edition: six cards in one boot via sequential driver_override chain, per-boot re-unlock runbook, Gen3/width findings. Fork of NatsumeAi's original.

   Shell 2

5. youtube-dl-to-telegram youtube-dl-to-telegram Public

   Python 1

6. llama-cpp-cmp90hx llama-cpp-cmp90hx Public

   llama.cpp fork with CMP 90HX optimizations: DP4A→IMAD, FFMA→HFMA2 (+58% decode speed)

   C++ 1