Topics / #avx2

Avx2 projects on GitHub

SIMD optimization using AVX2 vector instructions for high-performance code.

As of , the most-starred Avx2 projects on GitHub are simdjson (24.5K stars), kimi-k3-in-c (9.3K stars) and LASRecomp (2.1K stars). GitHub.Locker tracks 5 Avx2 repositories with 36.5K stars combined. The fastest-growing on October 31 was xcx-recomp, with +72 stars in a day.

5Projects
36.5KStars in total
+149Stars in a day
Browse in the catalog →github.com/topics/avx2 ↗

Hot right now

Most stars gained in a day · Oct 31

+72

xcx-recomp

jmars125/xcx-recomp · ★ 120

XCX Recomp statically recompiles Xenoblade Chronicles X: Definitive Edition (Nintendo Switch, 2025) into a native Windows x86-64 application with a Vulkan renderer, so the game runs on PC without an emulator.

+29

LASRecomp

peiikai/LASRecomp · ★ 2.1K

LAS Recomp statically recompiles The Legend of Zelda: Link's Awakening (Nintendo Switch, 2019) into a native application for Windows, macOS and Linux, with a built-in randomizer.

+28

janas

prabanta-dev/janas · ★ 566

Janas-LLM is a C-based inference engine that runs large mixture-of-experts LLMs (even ones bigger than your RAM) on ordinary laptops, by streaming the experts each token needs from an NVMe SSD and caching the hot ones in memory.

+15

kimi-k3-in-c

FareedKhan-dev/kimi-k3-in-c · ★ 9.3K

An inference engine for the 2.78-trillion-parameter Kimi K3 model, written in portable C99. It runs on a plain CPU in as little as 8 GB of RAM — no GPU, no BLAS, no framework — streaming the 1.56 TB checkpoint straight off disk.

The all-time top

5 projects · by stars

  1. #1 · simdjson/simdjson

    simdjson

    A C++ library that parses JSON at gigabytes-per-second speed using SIMD instructions and microparallel algorithms. It ships in production systems like Node.js, ClickHouse and Meta Velox.

    24.5K
  2. #2 · FareedKhan-dev/kimi-k3-in-c

    kimi-k3-in-c

    An inference engine for the 2.78-trillion-parameter Kimi K3 model, written in portable C99. It runs on a plain CPU in as little as 8 GB of RAM — no GPU, no BLAS, no framework — streaming the 1.56 TB checkpoint straight off disk.

    9.3K
  3. #3 · peiikai/LASRecomp

    LASRecomp

    LAS Recomp statically recompiles The Legend of Zelda: Link's Awakening (Nintendo Switch, 2019) into a native application for Windows, macOS and Linux, with a built-in randomizer.

    2.1K
  1. 4

    janas

    566prabanta-dev/janas · C

    Janas-LLM is a C-based inference engine that runs large mixture-of-experts LLMs (even ones bigger than your RAM) on ordinary laptops, by streaming the experts each token needs from an NVMe SSD and caching the hot ones in memory.

  2. 5

    xcx-recomp

    120jmars125/xcx-recomp · C

    XCX Recomp statically recompiles Xenoblade Chronicles X: Definitive Edition (Nintendo Switch, 2025) into a native Windows x86-64 application with a Vulkan renderer, so the game runs on PC without an emulator.

Often tagged together