This course covers concepts and approaches related to programming GPU processors using OpenACC directives and the PGI Accelerator programming model. Extensive coverage of GPU hardware, memories, data transport, and performance optimization enable the student to understand the fundamental aspects of GPU programming. In-depth, hands-on lectures and laboratories demonstrate how to apply OpenACC directives to serial software. Using a directive based approach, students will capitalize on low-cost, high performance GPU computing hardware to improve application performance while reducing maintenance and porting requirements.
Length: 2 Days; Cost: $1895
What You’ll Learn
- Correctly indentify concurrency opportunities and parallelize algorithms to run on the GPU.
- Install NVIDIA and PGI tools and compile CUDA and OpenACC programs.
- Understand the NVIDIA GPU hardware platform and the underlying technical architecture, including high-throughput SIMD processing and hardware threading architecture concepts.
- Recognize the difference between GPU memory types and the advantages and disadvantages of each.
- Learn to determine the best methods for software development with the OpenACC API.
- Understand the PGI Accelerator command set and its application to C and Fortran codes.
- Learn the specific skills to accelerate applications on x64+GPU platforms with the PGI Accelerator compilers.
- Learn to tune data movement, memory loads and stores, and loop schedules for maximum effect.
- Effectively orchestrate the tranport of data to and from GPU memory.
- Learn to meld multicore processors and GPUs to take maximum advantage of modern platform performance.
- Discover how to take advantage of multiple GPUs in the same computer.
- Cover debugging strategies for PGI Accelerator codes.
- Participate in hands-on laboratories to reinforce the theory and concepts presented in the class.
[Submitted by Ian Lintault, nCore Design]
Category: Training & Events