simdjson
Parsing gigabytes of JSON per second : used by Facebook/Meta Velox, the Node.js runtime, ClickHouse, WatermelonDB, Apache Doris, Milvus, StarRocks
- 25k
- C++
- Apache-2.0
- 3 d ago
Parsing gigabytes of JSON per second : used by Facebook/Meta Velox, the Node.js runtime, ClickHouse, WatermelonDB, Apache Doris, Milvus, StarRocks
A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.
SIMD, SWAR & GPU kernels for strings: search, hashing, sorting, edit distances, UTF-8 segmentation & ciphers for C, C++, CUDA, Python, Rust, Go, Swift, JS, C# & Java 🦖
High-efficiency floating-point neural network inference operators for mobile, server, and Web
Mixed-precision math, linear-algebra & retrieval kernels, 4000+ symbols from 4-bit numbers to 128-bit complex, for x86, Arm, RISC-V, LoongArch, Power & WASM, AMX & SME included 📐
37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.
Library for specialized dense and sparse matrix operations, and deep learning primitives.
Portable wrapper for SIMD and vector instructions written in C++11. Compatible with NEON, SSE, AVX, AVX-512 and SVE (length specific).
Pure-Rust, CPU-only OCR engine for Baidu Unlimited-OCR (a DeepSeek-OCR-derived 3B MoE VLM). Five-model zoo, custom int8 kernels, no ML framework, no Python, no GPU.
Neural networks and weight-space programming in Rust, JavaScript and WebAssembly. Inspect activations, edit weights, evaluate candidates, train small models, and explicitly promote or roll back revisions.
A Java version of simdjson, a high-performance JSON parser utilizing SIMD instructions
The HPC toolbox: fused matrix multiplication, convolution, data-parallel strided tensor primitives, OpenMP facilities, SIMD, JIT Assembler, CPU detection, state-of-the-art vectorized BLAS for floats and integers
Official library for RaBitQ vector quantization and approximate nearest-neighbor search, in C++ and Python.
Rust edit distance routines accelerated using SIMD. Supports fast Hamming, Levenshtein, restricted Damerau-Levenshtein, etc. distance calculations and string search.
The explainable, local-first memory engine for AI agents. One ~9 MB binary fuses vector + graph + columnar under VelesQL; why() returns the evidence path behind every recall. No cloud, no glue code — runs on server, browser, mobile and desktop.
SIMD-based linear algebra and statistics for data science with dart
World's First NMS-Free YOLOv26n on ESP32-P4. Features end-to-end Int8 QAT and custom C++ optimizations achieving 30% faster inference than the official ESP-DL YOLOv11n (1.7s vs 2.4s).
Permissive UTF-8/16/32 transcoding, comparison and BOM detection with SIMD ASCII fast paths
High-performance quantum computing in pure Rust + WebAssembly: state-vector circuit simulator with SIMD & noise models, VQE, Grover, QAOA, surface-code error correction, and real-time coherence via dynamic min-cut.
Rust port of libjpeg-turbo with equivalent or better performance