This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
SSE intrinsics + auto-alignment (3.2.2)
- From: ozzy at orkysquad dot org
- To: gcc at gnu dot org
- Date: Mon, 10 Feb 2003 12:19:41 +0100 (MET)
- Subject: SSE intrinsics + auto-alignment (3.2.2)
Heya dudes,
I just would like to comment here about the way __m128 and derived data types
from the
SSE intrinsics part of the new gcc3.2.2 fix release are handled.. I'm
currently working
(@orkysquad.org) on a game project which should use intrinsics. Unfortunately,
when
came the time for a Linux port i get one of my worst headache so far (since i
first met
DirectX! ;).
My point is, for portability purposes it is a really bad option to keep SSE
intrinsics data
types *not* auto aligned (into structures) while Icc and Msvc compilers are
doing that (and
maybe __int64 aswell but can't remember). So, it is not difficult to port your
intrinsics code
to GCC but really really difficult to port your win32 defined structures
(data)!! That's too
bad why don't everybody uses the Intel(1998 guidelines) and mxxxxxxxxt way of
handling
padding for specific SSE intrinsics?? frankly it will be a better world for
everyone! ;))
Thus, Is it planned to correct this issue for GCC3.3 release and above? In the
negative, why
and what does it involve? :-p
Thank u!
tchOo!
Ozzy.
ps: (is there any forum btw just like gl forums @opengl.org for gcc purposes?)