Introduction to GPU Programming Training Course

Introduction

• Defining GPU programming
• Reasons for adopting GPU programming
• Challenges and trade-offs in GPU development
• Overview of available frameworks and tools
• Selecting the optimal framework and tool for your application

OpenCL

• Understanding OpenCL
• Advantages and disadvantages of OpenCL
• Configuring the OpenCL development environment
• Developing a basic OpenCL program for vector addition
• Utilizing the OpenCL API to query device information, manage memory, transfer data, launch kernels, and synchronize threads
• Writing device-executable kernels using the OpenCL C language
• Using OpenCL built-in functions, variables, and libraries for common operations
• Optimizing data transfers and memory access via OpenCL memory spaces (global, local, constant, private)
• Controlling parallelism using the OpenCL execution model (work-items, work-groups, ND-ranges)
• Debugging and testing OpenCL programs with tools like CodeXL
• Optimizing OpenCL programs through coalescing, caching, prefetching, and profiling

CUDA

• Understanding CUDA
• Advantages and disadvantages of CUDA
• Setting up the CUDA development environment
• Creating a basic CUDA program for vector addition
• Using the CUDA API to query device information, manage memory, copy data, launch kernels, and synchronize threads
• Writing kernels in CUDA C/C++ to execute on the device and manipulate data
• Utilizing CUDA built-in functions, variables, and libraries for standard tasks
• Optimizing memory access using CUDA memory spaces (global, shared, constant, local)
• Managing parallelism via the CUDA execution model (threads, blocks, grids)
• Debugging and testing CUDA programs using CUDA-GDB, CUDA-MEMCHECK, and NVIDIA Nsight
• Optimizing CUDA programs using techniques such as coalescing, caching, prefetching, and profiling

ROCm

• Understanding ROCm
• Advantages and disadvantages of ROCm
• Configuring the ROCm development environment
• Developing a basic ROCm program for vector addition
• Using the ROCm API to query device information, manage memory, transfer data, launch kernels, and synchronize threads
• Writing kernels in ROCm C/C++ to execute on the device
• Using ROCm built-in functions, variables, and libraries for common operations
• Optimizing memory access using ROCm memory spaces (global, local, constant, private)
• Controlling parallelism via the ROCm execution model (threads, blocks, grids)
• Debugging and testing ROCm programs using the ROCm Debugger and ROCm Profiler
• Optimizing ROCm programs through coalescing, caching, prefetching, and profiling

HIP

• Understanding HIP
• Advantages and disadvantages of HIP
• Setting up the HIP development environment
• Creating a basic HIP program for vector addition
• Writing kernels in HIP to execute on the device and manipulate data
• Using HIP built-in functions, variables, and libraries for standard tasks
• Optimizing memory access using HIP memory spaces (global, shared, constant, local)
• Managing parallelism via the HIP execution model (threads, blocks, grids)
• Debugging and testing HIP programs using the ROCm Debugger and ROCm Profiler
• Optimizing HIP programs using techniques such as coalescing, caching, prefetching, and profiling

Comparison

• Comparing features, performance, and compatibility across OpenCL, CUDA, ROCm, and HIP
• Evaluating GPU program efficiency using benchmarks and metrics
• Adopting best practices and tips for GPU programming
• Exploring current trends and future challenges in GPU programming

Summary and Next Steps