Uploaded image for project: 'JDK'
  1. JDK
  2. JDK-8192846

Support cmov vectorization for float

    XMLWordPrintable

    Details

    • Type: Enhancement
    • Status: Resolved
    • Priority: P3
    • Resolution: Fixed
    • Affects Version/s: 9
    • Fix Version/s: 10
    • Component/s: hotspot
    • Labels:
    • Subcomponent:
    • Resolved In Build:
      b36
    • CPU:
      x86_64
    • OS:
      linux

      Backports

        Description

        JDK-8139340 has added vectorization support for double cmoves. Float support is a trivial extension which can piggyback on top of existing algorithm.

        With such an enhancement, the following pattern can be supported:
        {code}
        private void cmove_kernel_float(float[] in1, float[] in2, int length, float[] out) {
          for (int i = 0; i < length; i++) {
            out[i] = (in1[i] > in2[i]) ? in1[i] : in2[i];
          }
        }
        {code}

          Attachments

            Issue Links

              Activity

                People

                Assignee:
                rlupusoru Razvan Lupusoru
                Reporter:
                rlupusoru Razvan Lupusoru
                Votes:
                0 Vote for this issue
                Watchers:
                5 Start watching this issue

                  Dates

                  Created:
                  Updated:
                  Resolved: