writes:
> On Wed, 2009-09-02 at 09:05 -0700, Keith Whitwell wrote:
> > On Wed, 2009-09-02 at 08:48 -0700, José Fonseca wrote:
> > > On Wed, 2009-09-02 at 05:11 -0700, Chris Wilson wrote:
> > > > By rearranging the bitfields within the key we can reduce the size
> > > > of the key from 644 to 196 bytes, reducing the cost of both the
> > > > hashing and equality tests.
> > > > ---
> > > > src/mesa/main/texenvprogram.c | 7 ++++---
> > > > 1 files changed, 4 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/src/mesa/main/texenvprogram.c b/src/mesa/main/texenvprogra
> m.c
> > > > index 5913957..3851937 100644
> > > > --- a/src/mesa/main/texenvprogram.c
> > > > +++ b/src/mesa/main/texenvprogram.c
> > > > @@ -82,8 +82,8 @@ texenv_doing_secondary_color(GLcontext *ctx)
> > > > #define DISASSEM (MESA_VERBOSE & VERBOSE_DISASSEM)
> > > >
> > > > struct mode_opt {
> > > > - GLuint Source:4; /**< SRC_x */
> > > > - GLuint Operand:3; /**< OPR_x */
> > > > + GLubyte Source:4; /**< SRC_x */
> > > > + GLubyte Operand:3; /**< OPR_x */
> > > > };
> > >
> > > This doesn't compact anything. This chunk stay as is.
> >
> > The size of the struct decreases from 4 bytes to one.
[snip -- simplified example]
> Interesting. Gcc is trying to do something smart with alignment.
In that case, wouldn't adding
__attribute__((__packed__))
to the struct def have the same effect (and not rely on a side effect
of how gcc's alignment algorithm works)?
I'm not sure of the MSVC equivalent, but I imagine an equivalent
exists.
-tom
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Mesa3d-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mesa3d-dev
