This topic explores GPU architecture and CUDA programming, focusing on three key aspects. First, it discusses the SIMT (Single Instruction, Multiple Threads) block-based architecture and how CUDA kernels leverage this parallelism. Second, it covers memory, for efficient data access. Finally, it examines hardware and software synchronization methods, such as barriers, essential for coordinating thread execution in parallel computing.

An introduction to the 3D graphics pipeline used by GPUs in personal computers. It should walk the reader through the main stages involved in transforming raw geometric data into pixels displayed on screen.

This article goes over the evolution of the hardware graphics pipeline on GPUs across the different game console generations.

This article goes over the evolution of the hardware graphics pipeline on GPUs across the different game console generations.