2010/9/15 Ozkan Sezer <[email protected]>
> Hrmph, the compile command line doesn't have any -O options, meaning
> that __OPTIMIZE__ is not defined. See xmmintrin.h provided by gcc:
>
> #ifdef __OPTIMIZE__
> extern __inline __m64 __attribute__((__gnu_inline__,
> __always_inline__, __artificial__))
> _mm_insert_pi16 (__m64 const __A, int const __D, int const __N)
> {
> return (__m64) __builtin_ia32_vec_set_v4hi ((__v4hi)__A, __D, __N);
> }
>
> extern __inline __m64 __attribute__((__gnu_inline__,
> __always_inline__, __artificial__))
> _m_pinsrw (__m64 const __A, int const __D, int const __N)
> {
> return _mm_insert_pi16 (__A, __D, __N);
> }
> #else
> #define _mm_insert_pi16(A, D, N) \
> ((__m64) __builtin_ia32_vec_set_v4hi ((__v4hi)(__m64)(A), \
> (int)(D), (int)(N)))
>
> #define _m_pinsrw(A, D, N) _mm_insert_pi16(A, D, N)
> #endif
>
> Since __OPTIMIZE__ is not defined, the macro version is in effect and it is
> messing up things. In the end, lines 510 and 517 of intrin.h, which are :
>
> __MACHINEX86X_NOX64(__m64 _m_pinsrw(__m64,int,int))
>
> ... and
> __MACHINEX86X_NOX64(__m64 _m_pshufw(__m64,int))
>
> ... get translated into:
> __m64 ((__m64) __builtin_ia32_vec_set_v4hi ((__v4hi)(__m64)(__m64),
> (int)(int), (int)(int)));
>
> ... and:
> __m64 ((__m64) __builtin_ia32_pshufw ((__v4hi)(__m64)(__m64), (int)(int)));
>
> ... and the compiler rightfully screams.
>
> Adding Kai to the CC list.
>
> --
> O.S.
>
Should I add -O2 manually to CFLAGS when running configure as a workaround?
Ruben
------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev
_______________________________________________
Mingw-w64-public mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
