New in Beta Update 5
Added the support of allowing separated worksharing for loop in parallel region for offloading. Level-Zero based OpenMP* offloading support. Added static / global support for declare target. Improved reduction performance by leveraging Gen SLM. Enhance hierarchical parallelism support for more complicated cases in target region. Fortran offload support from an interface block in a module. Improved LLVM-Based F90 compiler quality and added more functionalities.