The LLVM/Clang compiler toolchain received updates yesterday to enhance its support for AMD Zen 1, Zen+, and Zen 2 processors. This work, contributed by the open-source community, specifically addresses instruction scheduling for PDEP and PEXT operations. These particular instructions have been noted as "quite costly" for the original Zen processors, leading to performance bottlenecks. The improvements provide more informed instruction scheduling, aiming to mitigate the performance overhead associated with these operations. This change is expected to yield better overall performance for applications compiled with LLVM/Clang on the affected Zen 1, Zen+, and Zen 2 architectures.