07-28, 16:40–16:50 (UTC), Purple
In this talk, updates on the development of a GPU backend for Apple hardware (specifically the M-series chipset) will be presented along with a brief showcase of current capabilities and interface. The novel compilation flow will be explained and compared to the other GPU backends as well as the benefits and limitations of both a unified memory model and Apple's Metal capabilities. A brief overview of Apple's non-GPU hardware accelerators and their potential will also be discussed.
The release of Apple's M-series chipset brings new hardware into play for Julia to target. Base CPU functionality is already highly used within the community, but so far, the M1 chip's hardware accelerators have primarily been inaccessible to Julia programmers. Metal.jl has been developed as a GPU backend (like CUDA./, AMD.jl, and oneAPI.jl) specifically targeting the M-series GPUs. Given Apple's continued expansion of the M1 chipset and devotion to hardware accelerators, a Julia interface targeting these compute devices is becoming increasingly beneficial.
Tim Besard is a software engineer at Julia Computing, working on GPU support for the Julia language. He holds a Ph.D. in computer science engineering from Ghent University, Belgium, for research on abstractions to program hardware accelerators in high-level programming languages.
Max is an undergraduate Computer Engineering student at the University of Alabama interested in helping scientists easily and effectively utilize their computing hardware.