Articles by tag: CUDA

NVIDIA CUDA cores accelerated by AI by 38% without human involvement

Multi-agent Cursor system autonomously optimized 235 CUDA cores for Blackwell B200. Results, methodology, and significance for HPC developers.

ROCm is catching up to CUDA in AI: OneROCm and Triton

AMD unified ROCm into OneROCm for all accelerators. Triton writes kernels for AMD/Nvidia. vLLM inference on par with CUDA. Six-week releases. Test the stack on Strix Halo.

SIMD optimization of Mandelbrot AVX2 OpenMP CUDA

Boosting the Mandelbrot set: from 7 FPS scalar C++ to 500+ FPS on CUDA. AVX2 intrinsics, OpenMP multithreading, GPU code. Benchmarks Ryzen 5 + RTX 3050. For middle/senior developers.

GameDev News: UNIGINE 2.21 and RTX Neural Shaders

Layoffs at Eidos and Take-Two, UNIGINE SDK 2.21 release, RTX Neural Shaders techniques and GPU acceleration. Technical details for developers. Check out reviews and behind-the-scenes.

GPU acceleration of the Game of Life: CUDA Triton benchmarks

Comparison of PyTorch, CUDA, Triton for the cellular automaton "Game of Life" on A40. Results 22.5 ms, kernel code. Study optimizations for memory-bound tasks. Read the analysis.