flame/how-to-optimize-gemm resource

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

2035 stars · 366 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A tutorial wiki repository by Prof. Robert van de Geijn teaching step-by-step how to optimize general matrix-matrix multiplication (GEMM) using the GotoBLAS/BLIS approach. It contains C code examples progressing from naive implementations to blocked, packed, cache-optimized kernels.

Use cases

• learn how to optimize matrix multiplication in C
• understand the GotoBLAS/BLIS approach to GEMM optimization
• study cache blocking and memory packing for dense linear algebra
• teach high-performance computing with step-by-step GEMM examples
• improve performance of my BLAS-like kernel code

When to choose

• you want to learn CPU-level optimization techniques like blocking, packing, and register tiling for matrix multiplication
• you are teaching or studying high-performance dense linear algebra
• you want to understand how BLAS libraries achieve near-peak GEMM performance

When to avoid

• you need a production-ready, optimized BLAS library - use OpenBLAS, BLIS, or MKL instead
• you are optimizing GEMM for GPUs rather than CPUs
• you want a maintained software library with a stable API and license

Facets

learning-resource · maturity maintenance

benchmarking developer-tools performance tutorials machine-learning cpp windows gemm matrix-multiplication gotoblas blis code-optimization high-performance-computing tutorial algorithms linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="flame/how-to-optimize-gemm")