John Ciliberti·3d agoBuilding a High-Performance Rust Compression Library with Spec-Driven Development and Constrained…Subtitle: Beating LZ4, Zstd, and Pigz on decompression speeds while exploring GPU acceleration, SIMD micro-optimizations, and autonomous…
InGoPenAIbyChirav Dave·Aug 26GPUs, CUDA, Drivers and PyTorch: Concept, Setup & InstallationThis article is intended to build a mental model of the complete GPU software stack so that installing and troubleshooting PyTorch on any…
Joshi Rutvik·Jul 24🚀 Local GPU Acceleration from Scratch: The Complete Guide to CUDA, PyTorch CNNs, and Real-World…Developing Deep Learning models in cloud environments is convenient, but nothing matches the speed, control, and lack of subscription fees…
InCommit to Code QualitybyRicardo M Camacho·Jun 22Bringing Rigorous Static Analysis to CUDA CodeAs AI models move from the cloud to the edge, powering autonomous vehicles, advanced robotics, and real-time medical diagnostics…
Pranjali Puranik·Jan 6Inside the GPU SM: Understanding CUDA Thread ExecutionSM (Streaming multiprocessor) on GPU is analogous to the core on CPU. While CPU cores are optimized for low-latency execution of a few…
Pranjali Puranik·Dec 31, 2025GPU Memory model — Part 2This diagram provides a high-level view of the GPU memory spaces we’ll discuss and how they relate to one another.
Pranjali Puranik·Dec 13, 2025CUDA Runtime vs Driver API — The Mental Model That Actually MattersThe Mental Model That Actually Matters
Edward Morgan·Dec 6, 2025cuTile Explained with a Restaurant AnalogyTraditional CUDA makes you think about thousands of GPU threads. You decide exactly which thread handles which element. It feels like…
