Projekte
Auf GitHub ansehen
Maharshi-Pandya/cudacodes
Maharshi-Pandya/cudacodes has 434 GitHub stars and 20 forks. GitHot measured -1 stars in 24 hours. Last checked 2026-11-19 17:20 UTC. Last push 2025-06-29.
Gemessener Sternzuwachs
- in 24h
- -1
- in 7 days
- Verlauf wird gesammelt
- in 15 days
- Verlauf wird gesammelt
- in 1 month
- Verlauf wird gesammelt
Nettoänderungen in überlappenden Zeiträumen, keine Tagessummen. Fehlende Messwerte bleiben leer. Sterne zeigen Interesse, nicht Qualität.
Über uns
Learnings and programs related to CUDA
LernenREADME.md
OriginalCUDA Code Explorations
This repository showcases my journey learning and experimenting with CUDA, primarily using C/C++. Inside, you'll find code examples, insights, and optimizations related to various CUDA concepts. Each directory focuses on a specific topic, with extensively commented code to guide understanding.
What's Inside?
This repo contains:
- Practical Examples: Hands-on CUDA code for common operations.
- Optimizations: Explores different optimization techniques to improve performance.
- Well-Commented Code: Every line is explained, making it easy to follow along.
- Learning Resource: A place to learn from and improve your own CUDA skills.
Quick Navigation
softmax/: Explores various softmax implementations, from naive to optimized.matmul/: Demonstrates matrix multiplication with tiling strategies.matvec/: Iteratively optimizes matrix-vector multiplication (SGEMV) to achieve cuBLAS-like performance.query-device/: A simple tool to get device information of GPU.flash-attention/: Exploration of flash attention algorithms (Work in progress).
Getting Started
- Clone the repository:
git clone https://github.com/Maharshi-Pandya/cudacodes.git
- Navigate to the directory of interest (e.g.,
cd matmul) - Check the
README.mdinside the directory to learn how to compile and run the examples.