BEATING NUMPY'S MATRIX MULTIPLICATION IN 150 LINES OF C CODE

Summary from the TLDR Web Dev newsletter

This author presents an optimized matrix multiplication implementation in C that follows the BLIS design which outperforms NumPy's OpenBLAS implementation on an AMD Ryzen 7700 processor by efficiently utilizing SIMD instructions and multi-threading.