GPU Programming with OpenCL Training Course

Introduction

• Defining OpenCL and its core principles.
• Comparing OpenCL with CUDA and SYCL.
• An overview of key OpenCL features and architectural components.
• Steps to set up the development environment.

Getting Started

• Initiating a new OpenCL project in Visual Studio Code.
• Navigating the project structure and identifying key files.
• Compiling and executing the initial program.
• Displaying output using standard functions like printf and fprintf.

OpenCL API

• The role of the OpenCL API within the host program.
• Querying device information and capabilities via the API.
• Creating and managing contexts, command queues, buffers, kernels, and events.
• Enqueueing commands for reading, writing, copying, mapping, unmapping, execution, and synchronization.
• Effective error and exception handling strategies.

OpenCL C

• The function of OpenCL C in device-side programming.
• Authoring kernels to execute logic and manipulate data on the device.
• Utilizing data types, qualifiers, operators, and expressions in OpenCL C.
• Applying built-in mathematical, geometric, and relational functions.
• Leveraging extensions and libraries, including atomic operations, image processing, and cl_khr_fp16.

OpenCL Memory Model

• Distinguishing between host and device memory architectures.
• Working with memory spaces: global, local, constant, and private.
• Managing memory objects such as buffers, images, and pipes.
• Configuring memory access modes, including read-only, write-only, and read-write.
• Implementing memory consistency and synchronization mechanisms.

OpenCL Execution Model

• Comparing host versus device execution paradigms.
• Defining parallelism through work-items, work-groups, and ND-ranges.
• Using work-item functions like get_global_id, get_local_id, and get_group_id.
• Applying work-group functions such as barrier, work_group_reduce, and work_group_scan.
• Querying device parameters with functions like get_num_groups, get_global_size, and get_local_size.

Debugging

• Identifying common errors and bugs in OpenCL applications.
• Inspecting variables, breakpoints, and call stacks using the Visual Studio Code debugger.
• Analyzing and debugging OpenCL programs on AMD devices with CodeXL.
• Debugging and analyzing OpenCL performance on Intel devices using Intel VTune.
• Debugging and analyzing OpenCL applications on NVIDIA devices with NVIDIA Nsight.

Optimization

• Factors influencing OpenCL program performance.
• Enhancing arithmetic throughput using vector data types and vectorization.
• Reducing control overhead and improving locality through loop unrolling and tiling.
• Optimizing memory bandwidth using local memory and associated functions.
• Measuring and improving execution time and resource utilization via profiling tools.