#cuda
Cuda: 5 AI articles covering cuda news, analysis, and research
Articles
Turning Conv2D Into a Practical GPU Compute Patternβ10
Explore how Conv2D becomes a practical GPU compute pattern, covering data reuse, memory hierarchy, parallelism, and mapping convolution loops to GPU hardware.
Meet qwenGPU: Qwen3.5-2B With 45 Compiled GPU Layers That Runβ7
qwenGPU compiles 45 GPU layers into Qwen3.5-2B, running real CUDA inside the model's forward passesβ69 layers total, tested exact on Apple, NVIDIA, and CPU.
NVIDIA Nsight Python Tutorial: How to Profile GPU Kernels withβ9
Profile NVIDIA GPU kernels in Python with Nsight Python and pandas DataFrames: annotate code, collect hardware metrics, sweep sizes, and compare the CUPTI backe...
Amazon Takes Its Biggest Shot at Nvidia Yet with New TorchNeuronβ10
AWS launches TorchNeuron, letting AI startups run PyTorch apps on Trainium with a one-line change, directly challenging Nvidia's CUDA software advantage.
How to Use NVIDIA Warp and MjWarp to Accelerate Roboticsβ10
NVIDIA Warp and MjWarp bring GPU-accelerated, differentiable robotics simulation to Python. Learn how to build fast physics and learning workflows.