Hacker GitHub

← All repos

warpfront/hipfire

View on GitHub ↗ ★ 810 Rust

RDNA-native LLM inference engine in Rust.

3 mentions on Hacker News by 1 person: 0 stories, 3 comments · first 2026-10-01 · latest 2026-11-01

Mentions per year 3 mentions · busiest year 2026 (3)
20082026
AllStoriesComments NewestTopOldest
  1. I wonder how it compares to Hipfire: https://github.com/warpfront/hipfire
    comment by d3Xt3r · 2026-11-01 · in An inference server for AMD GPUs, in C++ and HIP
  2. …and q38rocm/HaloPFX[2], they're so much more faster than llama.cpp for consumer AMD GPUs. [1] https://github.com/warpfront/hipfire [2] https://github.com/julianmb/q38rocm
    comment by d3Xt3r · 2026-10-13 · in Janas: Run LLMs larger than RAM by streaming experts from SSD
  3. …or HaloPFX[2]? Both are far superior to llama.cpp in terms of performance, for Strix Halo. [1] https://github.com/warpfront/hipfire [2] https://github.com/julianmb/halofpx
    comment by d3Xt3r · 2026-10-01 · in Gemini 4 Argon