[llvm-bugs] [Bug 29083] New: [X86][SSE] Improve costs for floating point operations

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=29083

            Bug ID: 29083
           Summary: [X86][SSE] Improve costs for floating point operations
           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: llvm-bugs at lists.llvm.org
    Classification: Unclassified

Having a base (throughput) cost of 2 for fadd/fsub/fmul/fdiv/fsqrt is just
wrong - either a little too high or phenomenally too low.

The scalar costs probably need reviewing as well.

Hopefully we can use SSE41/AVX/AVX2 as a reasonable analogue for the improved
throughputs of FDIV/FSQRT with more recent hardware (SB, HSW, SKL).

-- 
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/20160821/2503f683/attachment.html>