This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RFC: A new meta intrinsic header file for x86 intrinsics


On Fri, Apr 4, 2008 at 10:58 AM, Chris Lattner <clattner@apple.com> wrote:
>
>
>  On Apr 4, 2008, at 10:13 AM, H.J. Lu wrote:
>
>
> > Hi,
> >
> > For each new set of x86 intrinsics, we introduce a new header file. It
> > will be desirable for users just to include one header file for all
> > intrinsics, current and future. Icc has <ia32intrin.h>, which includes
> > proper individual intrinsic header files and users just need
> > to include <ia32intrin.h> to get access to all intrinsics.
> >
> > We will implement meta intrinsic header file for AVX intrinsics.
> > However, I am not sure if <ia32intrin.h> is the best name for it.
> > Does anyone have any suggestions, how about <x86intrin.h>?
> >
>
>  I prefer <x86intrin.h>, as these are presumably usable in x86-64 mode.
>
>  One random request: would it be possible to keep mm_malloc.h out of the
> umbrella header?  Inclusion of mm_malloc.h make use of SSE difficult in
> kernel contexts, as mm_malloc.h pulls in stdlib.h and errno.h.
>

The idea is one header file to make all intrinsics available, including
_mm_malloc () and _mm_free ().  Unless we exclude those 2, I
don't see that we have a choice.

H.J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]