2.1.Intro to CUDA C++

View original course

6 study materials · 4 sections

This course provides a comprehensive introduction to CUDA C++ programming, covering the fundamental architecture and programming model of NVIDIA GPUs. Students will learn how to leverage the Single Instruction, Multiple Threads (SIMT) model to write massively parallel kernels and manage complex memory hierarchies. The curriculum spans from basic kernel invocation and compilation with the nvcc toolchain to advanced topics like asynchronous execution with streams and unified memory management.

Course Sections

Introduction to CUDA C++ and the NVCC Compiler

Key concepts: __global__ kernels · Triple chevron notation (<<< >>>) · nvcc Toolchain · PTX and Cubin Generation · Compilation Workflow

Covers the basics of CUDA C++ programming, including kernel definition, the compilation workflow using nvcc, and the distinction between host and device code.

The SIMT Execution Model and Thread Hierarchy

Key concepts: SIMT Execution Model · Thread Blocks and Grids · threadIdx, blockIdx, blockDim, gridDim · Thread Indexing

Explains the Single Instruction, Multiple Threads (SIMT) model and how threads are organized into grids and blocks for parallel execution.

Memory Management and Unified Memory

Key concepts: Global Memory · Shared Memory · Unified Virtual Address Space (UVA) · Managed Memory · HMM and ATS

Explores the GPU memory hierarchy and Unified Memory features that simplify data management between the CPU and GPU.

Asynchronous Execution and Streams

Key concepts: CUDA Streams · CUDA Events · Asynchronous Execution · Pinned Memory · Blocking vs. Non-blocking

Covers techniques for overlapping computation and memory transfers to maximize GPU utilization using Streams and Events.

Source Materials

Study 2.1.Intro to CUDA C++ on Lykke

This course page is free to open: signed out, you can read its first two sections, and signing in, on any plan, opens the rest. The free Lykke Chrome extension syncs your own Canvas courses, so ChatGPT, Claude and Muse can answer from your classes and deadlines. It only reads Canvas.

Add the free extensionSign in to open every section

Browse more course pages