[llvm-bugs] [Bug 36931] New: [X86] Split WriteIMul into 8/16/32/64 implementations

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=36931

            Bug ID: 36931
           Summary: [X86] Split WriteIMul into 8/16/32/64 implementations
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: llvm-dev at redking.me.uk
                CC: a.bataev at hotmail.com, andrea.dibiagio at gmail.com,
                    clement.courbet at gmail.com, craig.topper at gmail.com,
                    llvm-bugs at lists.llvm.org, spatel+llvm at rotateright.com
            Blocks: 36550, 36908

MUL64 definitely needs its own scheduler class, possibly the others as well.
Multiply by constant (IMUL32rmi etc.) may need splitting off as well.

Note: Integer multiplies seem to be a common issue in vectorization regressions
(incorrectly comparing scalar/vector integer multiply costs), so all need to be
accurate for [Bug #36550] to work well.


Referenced Bugs:

https://bugs.llvm.org/show_bug.cgi?id=36550
[Bug 36550] Derive TTI instruction costs from scheduling models
https://bugs.llvm.org/show_bug.cgi?id=36908
[Bug 36908] [meta][x86] Improve scheduler classes instruction coverage
-- 
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/20180328/ba31d69a/attachment.html>