https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70421
--- Comment #1 from Zdenek Sojka <zsojka at seznam dot cz> --- The operation done by the vmovdqa32 instruction is inverted; this fixes the assembly (-O3, intel syntax): @@ -72,7 +72,7 @@ and rsp, -64 #, push QWORD PTR [r10-8] # push rbp # - mov eax, 2 # tmp108, + mov eax, 0xfd # tmp108, kmovw k1, eax # tmp108, tmp108 xor edx, ecx # tmp106, tmp100 .cfi_escape 0x10,0x6,0x2,0x76,0