- Re-organize vector API java implementation to use single C2 compiler intrinsic entry point for both masked and non-masked operations.
- Pass an explicit mask object (and associated class instance) to various vector intrinsic routines.
- Change existing vector API specific inline expansion routines (LibraryCallKit::inline_vector_*) to generate masked IR nodes for targets which support predicated vector instruction (SVE/AVX512) else generate appropriate graph shape using VectorBlend operation.