AOMP 22.0-1 Brings Many Improvements for AMD's Fortran Compiler GPU Offloading

AOMP 22.0-1 was released on Wednesday as the newest routine update to this downstream of LLVM/Clang/Flang maintained by AMD that continues to carry their latest modifications for enhancing the C/C++/Fortran compiler offloading support to AMD Radeon/Instinct hardware using the likes of OpenMP and OpenACC.

AOMP 22.0-1 marks their first release since re-basing to the latest LLVM Git upstream that is now tracking LLVM 22 development over the recent LLVM 21 stable release.

This open-source AMD downstream of LLVM focused on delivering the best GPU offloading support until the relevant patches are upstreamed is now even more capable for its Fortran support. AMD engineers contributed a number of patches to LLVM's Flang compiler with mapping improvements, reduction support in more cases, improved alias analysis, and various other enhancements for this Fortran compiler in the context of better offloading support.

AOMP 22.0-1 is also now built against the ROCm 7.0.1 source code, aomp-shell-format was merged, better build support for OpenCL, and various other changes.

The AOMP 22.0-1 source code as well as pre-built Ubuntu / Red Hat / SUSE Enterprise Linux binaries are available via GitHub.