ashvardanian/less_slow.cpp resource

Health v2 · maintenance only

Adoption not part of the score

1924 stars · 83 forks observed · 2026-08-28

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

A tutorial-style repository of benchmarks demonstrating performance-oriented coding practices in C++20, C, CUDA, PTX, and assembly, covering topics from SIMD numerics to coroutines, ranges, exceptions, and user-space IO. It uses Google Benchmark to build intuition for writing faster low-level code.

Use cases

• learn performance-oriented C++ coding practices
• benchmark SIMD vs scalar code
• understand CUDA and PTX micro-kernels
• measure coroutine and ranges overhead
• learn compiler optimization flags beyond -O3
• study branch predictor and memory behavior
• learn io_uring and user-space IO patterns

When to choose

• you want hands-on benchmarked examples of low-level C++/CUDA optimization
• you're building intuition for nanosecond-to-millisecond scale performance decisions
• you want a practical companion to HPC or systems programming study

When to avoid

• you need a production library to import into your project
• you want beginner-friendly C++ tutorials rather than performance deep-dives
• you need Windows or macOS-first guidance, since it targets GCC/Clang/NVCC on Linux

Facets

learning-resource · maturity active

benchmarking developer-tools math concurrency networking performance developer-tools tutorials gpu-computing cpp c cli cpp20 cuda simd assembly ptx coroutines io-uring google-benchmark hpc micro-optimization avx512 algorithms linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="ashvardanian/less_slow.cpp")