[LLVMbugs] [Bug 13336] New: ARM NEON intrinsic vshlq_n_u64 fail on shifts over 7.

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Wed Jul 11 14:26:55 PDT 2012


http://llvm.org/bugs/show_bug.cgi?id=13336

             Bug #: 13336
           Summary: ARM NEON intrinsic vshlq_n_u64 fail on shifts over 7.
           Product: clang
           Version: 3.1
          Platform: Other
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Headers
        AssignedTo: unassignedclangbugs at nondot.org
        ReportedBy: ted at krovetz.net
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified


Compiling

#include <arm_neon.h>
void foo()
{
    uint64x2_t x = {1,1};
    x = vshlq_n_u64(x,32);
}

gives the error "test.c:5:9: error: argument should be a value from 0 to 7"
indicating that a shift of 32 is illegal whereas

#include <arm_neon.h>
void foo()
{
    uint64x1_t x = {1};
    x = vshlq_n_u64(x,32);
}

works just fine. Both codes work as expected under GCC. Similar behavior is
seen when using vshrq_n_u64 too.

-- 
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