GPU Programming

  1. GPU Programming Models and APIs
    1. Overview of Programming Models
      1. Low-Level APIs
        1. NVIDIA CUDA
          1. OpenCL
            1. DirectCompute
              1. Vulkan Compute
                1. Metal Performance Shaders
                2. High-Level Frameworks
                  1. OpenACC
                    1. OpenMP Target Offloading
                      1. SYCL
                        1. Kokkos
                        2. Domain-Specific Languages
                          1. Halide
                            1. ArrayFire
                              1. Thrust
                            2. Choosing a Programming Model
                              1. Vendor Specificity vs. Portability
                                1. Hardware Compatibility
                                  1. Cross-platform Support
                                    1. Ecosystem Lock-in
                                    2. Performance Considerations
                                      1. Low-level vs. High-level Abstractions
                                        1. Optimization Opportunities
                                          1. Driver Overhead
                                          2. Development Factors
                                            1. Learning Curve
                                              1. Debugging Tools
                                                1. Community Support
                                                  1. Library Availability

                                              Previous

                                              1. Introduction to Parallel Computing and GPU Architecture

                                              Go to top

                                              Next

                                              3. Fundamentals of CUDA Programming