Efficient GPGPU programming
Efficient GPGPU programming, Ashot Vardanian In this talk, we will cover the differences between writing software for CPUs in C++, GPU kernels in OpenCL and CUDA, and heterogeneous C++ code. We will benchmark hand-written kernels and compare them to existing high-level libraries. To wrap things up, we will discuss compilers and the future of heterogeneous computing. The talk was presented at the distributed meetup between St. Petersburg C++ User Group and MUC++, Sep 5th, 2019.
Efficient GPGPU programming, Ashot Vardanian In this talk, we will cover the differences between writing software for CPUs in C++, GPU kernels in OpenCL and CUDA, and heterogeneous C++ code. We will benchmark hand-written kernels and compare them to existing high-level libraries. To wrap things up, we will discuss compilers and the future of heterogeneous computing. The talk was presented at the distributed meetup between St. Petersburg C++ User Group and MUC++, Sep 5th, 2019.