Modern CUDA Learn Notes with PyTorch for Beginners, 200+ CUDA Kernels, Tensor Cores, HGEMM, FA-2 MMA.

CUDA Core Compute Libraries

A curated list of best cuda programming books

μ-Cuda, COVER THE LAST MILE OF CUDA. With features: intellisense-friendly, structured launch, automatic cuda graph generation and updating.

A dynamic binary instrumentation tool for tracing and analyzing GPU kernel instructions.

Suckless no dependencies CUDA/C++ port of NVIDIA's DVLT, feed it images, get a 3D point cloud + camera poses. NO python. One fast 5MB binary.

Efficient implementations of Merge Sort and Bitonic Sort algorithms using CUDA for GPU parallel processing, resulting in accelerated sorting of large arrays. Includes both CPU and GPU versions, along with a performance comparison.

A beginner's guide to CUDA programming

Gouda is a CUDA transpiler. Take your CUDA C programs and recompile for OpenCL and more.

An implementation deepdive of scan algorithms in CUDA C++

test stream of cusolver

A C++ header-only library for parallel linear algebra on GPUs (CUDA/cuBLAS under the hood)

learning to develop lightning fast C++/CUDA neural network

CUDA C++ repository demonstrating advanced GPU computing, optimized parallel algorithms (FFTs, Tiled MatMul), and NVIDIA ecosystem integrations (cuBLAS, Thrust). Engineered for maximum throughput and HPC learning.

原生C++/CUDA/OpenBLAS实现的张量加速计算库。性能碾压torch（原生CUDA算子实现+0拷贝优化），且更加方便部署。

基于TensorN和CUDA算子加速的生成式预训练Transformer（GPT）语言模型的C++从零实现

Auditable, fail-closed, fuzz-tested GPU decompression in CUDA C++ — LZ4 batch-decoded on NVIDIA GPUs today; Snappy, GDeflate, and Zstd planned.

Reconstruct 3D point clouds from images using this zero-dependency CUDA inference engine for the DVLT model.

CUDA-based PPO reinforcement learning implementation built from scratch using a custom neural network . Extends CUDA n

A neural network implemented from scratch in C++/CUDA, including GPU forward propagation, backpropagation, gradient descent, memory management, and real-time training visualization.