Chapel is a programming language known for its effectiveness in parallel computing. The Chapel team has been working on adding GPU support to the language due to the increasing popularity of GPU computing. This tutorial introduces Chapel’s GPU programming features, assuming no prior knowledge of frameworks like CUDA and HIP. The tutorial covers the basics of using locales and the on statement to specify where code should be executed, as well as the use of order-independent loops for efficient GPU execution. The tutorial also mentions that Chapel’s GPU support is vendor-neutral, allowing code to be executed on both NVIDIA and AMD GPUs, or even on the CPU using the ‘CPU-as-device’ mode. Overall, the tutorial provides a beginner-friendly introduction to GPU programming in Chapel.
https://chapel-lang.org/blog/posts/intro-to-gpus/