https://bugs.llvm.org/show_bug.cgi?id=43453

            Bug ID: 43453
           Summary: Incorrect code generated with ext_vector_type(16)
           Product: clang
           Version: 9.0
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: C++
          Assignee: [email protected]
          Reporter: [email protected]
                CC: [email protected], [email protected],
                    [email protected], [email protected],
                    [email protected]

Consider the following code:

//--------------------------------------------------------------
typedef double Double16 __attribute__((ext_vector_type(16)));
// or
// typedef double Double16 __attribute__((vector_size(8*16)));

Double16 mult(const Double16& v1, const Double16& v2)
{
    return v1 * v2;
}
//--------------------------------------------------------------

When compiling with clang 9.0.0 using -Ofast -march=skylake, the following
result is produced:

//--------------------------------------------------------------
        vmovapd ymm0, ymmword ptr [rsi]
        vmovapd ymm1, ymmword ptr [rsi + 32]
        vmovapd ymm2, ymmword ptr [rsi + 64]
        vmovapd ymm3, ymmword ptr [rsi + 96]
        vmulpd  ymm0, ymm0, ymmword ptr [rdi]
        vmulpd  ymm1, ymm1, ymmword ptr [rdi + 32]
        vmulpd  ymm2, ymm2, ymmword ptr [rdi + 64]
        vmulpd  ymm3, ymm3, ymmword ptr [rdi + 96]
        ret
//--------------------------------------------------------------

I believe this is erroneous, as the result is not stored into the function
return value according to the ABI.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs

Reply via email to