AOMP 19.0-0 Released For AMD's OpenMP Offloading Compiler
AMD on Thursday published AOMP 19.0-0 as the newest version of their LLVM/Clang downstream compiler focused on delivering the latest OpenMP device offloading support for their Radeon GPUs and Instinct accelerators.
The version bump with AOMP 19.0-0 comes as a result of now tracking the latest LLVM/Clang 19.0 upstream Git given the recent stable release of LLVM/Clang 18. With this AOMP 19.0-0 release they are now building against the ROCm 6.0.2 sources, defaults now not to use multiple SDMA engines, performance improvements, ROCgdb support for the GFX1103 GPU, and various other fixes.
The AOMP 19.0-0 sources as well as binaries for Ubuntu, SUSE, and RHEL/CentOS systems can be obtained from GitHub.
The version bump with AOMP 19.0-0 comes as a result of now tracking the latest LLVM/Clang 19.0 upstream Git given the recent stable release of LLVM/Clang 18. With this AOMP 19.0-0 release they are now building against the ROCm 6.0.2 sources, defaults now not to use multiple SDMA engines, performance improvements, ROCgdb support for the GFX1103 GPU, and various other fixes.
The AOMP 19.0-0 sources as well as binaries for Ubuntu, SUSE, and RHEL/CentOS systems can be obtained from GitHub.
Add A Comment