[LLVMbugs] [Bug 6399] New: Nicer code for i32 vector multiplications

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=6399

           Summary: Nicer code for i32 vector multiplications
           Product: libraries
           Version: 2.6
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: enhancement
          Priority: P5
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: llvm at henning-thielemann.de
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4290)
 --> (http://llvm.org/bugs/attachment.cgi?id=4290)
Various LLVM implementations of integer vector multiplications

For 32 bit integer vector multiplications LLVM currently produces not so nice
code on SSE before SSE4.
On SSE4 pmulld is called, which is certainly the best we can have.
But on SSE2 we already have pmuludq which allows for several elegant
implementations.
Attached is a file containing functions that implement a certain vector
multiplication (suffix 'a') and several alternative implementations that
directly call pmuludq (suffix 'b').
The second file contains the assembly code produced by LLVM-2.6.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.