[llvm-bugs] [Bug 51972] New: Overly pessimistic cost for `trunc <16 x i32> -> <16 x i16>`

via llvm-bugs llvm-bugs at lists.llvm.org
Sat Sep 25 14:14:49 PDT 2021


https://bugs.llvm.org/show_bug.cgi?id=51972

            Bug ID: 51972
           Summary: Overly pessimistic cost for `trunc <16 x i32> -> <16 x
                    i16>`
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: lebedev.ri at gmail.com
                CC: craig.topper at gmail.com, llvm-bugs at lists.llvm.org,
                    llvm-dev at redking.me.uk, pengfei.wang at intel.com,
                    spatel+llvm at rotateright.com

Noticed when looking at LV -debug.

https://godbolt.org/z/vsWch848x
`trunc <8 x i32> -> <8 x i16>` has cost of 2,
but `trunc <16 x i32> -> <16 x i16>` has cost of 10.

Isn't that a bit too much?
Even for haswell it should be ~4.

-- 
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/20210925/15c84ed0/attachment.html>


More information about the llvm-bugs mailing list