Optimize long vector multiplication using x86 VPMUL[U]DQ instruction

XMLWordPrintable

    • Type: Enhancement
    • Resolution: Fixed
    • Priority: P4
    • 24
    • Affects Version/s: 24
    • Component/s: hotspot
    • b26
    • x86_64
    • generic

      Optimizes LongVector multiplication by emitting VPMULUDQ instruction for following patterns.

             MulVL ( AndV SRC1, 0xFFFFFFFF) ( AndV SRC2, 0xFFFFFFFF)
             MulVL (URShiftVL SRC1 , 32) (URShiftVL SRC2, 32)
             MulVL (URShiftVL SRC1 , 32) ( AndV SRC2, 0xFFFFFFFF)
             MulVL ( AndV SRC1, 0xFFFFFFFF) (URShiftVL SRC2 , 32)
             MulVL (VectorCastI2X SRC1) (VectorCastI2X SRC2)
             MulVL (RShiftVL SRC1 , 32) (RShiftVL SRC2, 32)

            Assignee:
            Jatin Bhateja
            Reporter:
            Jatin Bhateja
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: