tim612 writes:
>
> I use a standard sort of macro for invoking the rdtsc instruction from my
> C++ code, and it seems wo work fine until I turn on the "-O3" optimization
> flag. With this flag enabled, the values returned by this macro seem to be
> garbage.
>
> #define RDTSCLL(_x) __asm__ __volatile__ ("rdtsc" : "=A" (_x));
>
> My machine:
> Processor: AMD Opteron(tm) Processor 248
> Host type: x86_64-linux
> GCC version: gcc (GCC) 3.4.4 20050721 (Red Hat 3.4.4-2)
>
> Any ideas?
Can you supply a test case?
Andrew.
