CUDA (Compute Unified Device Architecture) is a parallel computing platform and programming model developed by NVIDIA for accelerating computation on NVIDIA GPUs (Graphics Processing Units). It provides a set of APIs, libraries, and development tools for writing parallel programs and executing them on GPU hardware for high-performance computing (HPC), scientific simulations, and deep learning applications. Readers can explore CUDA's programming model, architecture, and optimization techniques for harnessing GPU computing power and accelerating data processing and scientific computations, improving performance and scalability in GPU-accelerated applications.

CUDA

ZLUDA Has Been Seeing New Activity For CUDA On AMD GPUs

Tiny LLM hacks: Loading quantized model using Python/llama_cpp_python.

Announcing Confidential Computing General Access on NVIDIA H100 Tensor Core GPUs

The One Billion Row Challenge in CUDA: from 17m to 17s

llm.c

karpathy/llm.c: LLM training in simple, raw C/CUDA

AMD Releases Orochi 2.0 With More CUDA/HIP Functions Implemented For Better Portability

NVIDIA Wants More Programming Languages to Support CUDA

Efficient CUDA Debugging: Using NVIDIA Compute Sanitizer with NVIDIA Tools Extension and Creating Custom Tools

mlecauchois/micrograd-cuda