[LLVMbugs] [Bug 7303] New: Assert when comparing vectors on X86_64

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Fri Jun 4 09:59:36 PDT 2010


           Summary: Assert when comparing vectors on X86_64
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: ddneff at hotmail.com
                CC: llvmbugs at cs.uiuc.edu

I thought recent changes were supposed to allow comparison operations using
vectors on X86.  I get an assertion about 'Operand wasn't scalarized?' in
LegalizeTypes.h:510.  Here is a sample file that won't compile:

; ModuleID = 'bugoptasm.txt.bc'

@global_subject2_246951760 = internal global i64 0 ; <i64*> [#uses=1]
@global_subject11_246957600 = internal global i64 0 ; <i64*> [#uses=1]

define void @dependentGraph0() {

  %0 = load i64* @global_subject2_246951760      ; <i64> [#uses=1]
  %1 = load i64* @global_subject11_246957600     ; <i64> [#uses=1]
  %tupleInsert = insertelement <2 x i64> undef, i64 %0, i32 0 ; <<3 x i64>>
  %tupleInsert33 = insertelement <2 x i64> %tupleInsert, i64 %1, i32 1 ; <<3 x
i64>> [#uses=1]
  %q_equal35 = icmp eq <2 x i64> %tupleInsert33, zeroinitializer ; <<3 x i1>>
  call void @output_integer(i8* inttoptr (i64 246924032 to i8*), i64 0, i64 0,
<2 x i1> %q_equal35)
  ret void

declare void @output_integer(i8*, i64, i64, <2 x i1>) nounwind

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.

More information about the llvm-bugs mailing list