Performance Portability: From Fantasy to Reality
As we go through what many in high performance computing are calling a “Cambrian Explosion” of hardware architectures, what we would ideally like to be able to do is write software once and achieve maximum performance on all these new innovative platforms. We know that this ideal is not possible. However, we are making rapid progress towards achieving a level of performance portability that’s close enough to maximum performance while requiring relatively minimal effort to bring to new processor architectures. Managing this balance of productivity, performance and portability requires finding reasonable compromises between each. This talk is going to go through many of the most successful approaches and what trade-offs are required to deliver these into production. This talk will cover performance portability technologies using C++, SYCL, AI and various compiler technologies like MLIR.
Speaker(s)
Andrew Richards - Codeplay/CEO
CEO and co-founder of Codeplay, a pioneer in GPU acceleration, Andrew started his career writing video games in the days of 8-bit computers, progressing to write best-selling titles such as Pete Sampras Tennis. Andrew researched and developed early GPU compiler technology, and founded Codeplay in 2002. Codeplay have been producing compilers for games consoles, special-purpose processors and GPUs since then. Andrew chaired the SYCL standards group for its early years, which is now gaining rapid adoption, including now the C++ programming model for Intel’s oneAPI. Today, Codeplay is a world-leading specialist in heterogeneous computing technology, now working on extreme challenges from exascale to artificial intelligence and safety for self-driving cars. In July 2022, Codeplay became a subsidiary of Intel.