ajroetker/go-highway
Write SIMD once, run everywhere. AVX2, AVX-512, or pure Go fallback. Like Google Highway, but for Go.
This project helps Go developers accelerate their applications by processing large datasets faster. It allows you to write mathematical and data transformation operations once, and they will automatically run using specialized hardware instructions (like AVX2, AVX-512, or ARM NEON) if available, or fall back to standard Go code. It takes numerical data (like lists of numbers) and outputs processed data at significantly higher speeds, benefiting anyone working with Go for scientific computing, data analysis, or machine learning.
Use this if you are a Go developer building performance-critical applications that involve extensive numerical computations on arrays of data, especially for tasks like signal processing, image manipulation, or machine learning model inference.
Not ideal if your application does not involve heavy numerical array processing or if you are not a Go developer.
Stars
72
Forks
3
Language
Assembly
License
Apache-2.0
Category
Last pushed
Mar 09, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ai-coding/ajroetker/go-highway"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
abrander/gguf
GGUF parser for Go
thepabloaguilar/floc
A floc simulator wrapper over a Go implementation
heupr/heupr
The codebase for the Heupr application :tada:
theo-keyzer/ml-model-codegen
Translate high-level model definitions into highly-optimized, target-specific source code
velumlabs/kit
A Go/Rust dual-language package for creating and administering toolkits and tools for calling...