[llvm-bugs] [Bug 31223] New: [X86][AVX512] Should KSHIFT map to X86ISD::VSHLI and VSRLI ?

via llvm-bugs llvm-bugs at lists.llvm.org
Thu Dec 1 03:24:20 PST 2016


https://llvm.org/bugs/show_bug.cgi?id=31223

            Bug ID: 31223
           Summary: [X86][AVX512] Should KSHIFT map to X86ISD::VSHLI and
                    VSRLI ?
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: llvm-dev at redking.me.uk
                CC: craig.topper at gmail.com, elena.demikhovsky at intel.com,
                    llvm-bugs at lists.llvm.org
    Classification: Unclassified

X86ISD::VSHLI and VSRLI are for vector shifts within elements but kshift (<X x
i1> mask register types) uses it as a vector element shift (which is closer to
X86ISD::VSHLDQ and VSRLDQ).

So we see shifts such as X86ISD::VSHLI/VSRLI <16 x i1>, #8

This was encountered when I began looking at better VSHLI/VSRLI combining.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20161201/5a21b4c7/attachment.html>


More information about the llvm-bugs mailing list