CE9125 - Connect with the Experts: CUDA Fortran for NVIDIA GPUs

CUDA Fortran leverages high-level features of the Fortran language to present a GPU programming model that is easy to learn and use, has minimal impact on application portability, and enables the full flexibility and performance of CUDA. In this session you'll learn the basics of CUDA Fortran programming - managing the memory hierarchy with variable attributes, writing portable compute kernels with the !$CUF KERNELS directive, using Fortran generic interfaces to create portable code across CPUs and GPUs, using and writing GPU-accelerated libraries, using tensor cores on the V100. These topics and more will be presented, and attendees will come away with all the fundamentals they need to start using and writing CUDA Fortran programs using the free PGI Community Edition Fortran compiler.

Connect directly with NVIDIA engineers and experts from other organizations to get answers to all of your questions on topics ranging from AI & Deep Learning to Accelerated Data Science. Visit us at the pods in the exhibit hall and ask a question.