[Mesa-dev] [PATCH v5] mesa: use fi_type in vertex attribute code
marius.predut at intel.com
Tue Feb 24 03:19:42 PST 2015
> -----Original Message-----
> From: Brian Paul [mailto:brianp at vmware.com]
> Sent: Friday, February 20, 2015 5:49 PM
> To: Predut, Marius; mesa-dev at lists.freedesktop.org
> Subject: Re: [Mesa-dev] [PATCH v5] mesa: use fi_type in vertex attribute code
> On 02/19/2015 11:42 AM, marius.predut at intel.com wrote:
> > From: Marius Predut <marius.predut at intel.com>
> > For 32-bit builds, floating point operations use x86 FPU registers,
> > not SSE registers. If we're actually storing an integer in a float
> > variable, the value might get modified when written to memory. This
> > patch changes the VBO code to use the fi_type (float/int union) to
> > store/copy vertex attributes.
> > Also, this can improve performance on x86 because moving floats with
> > integer registers instead of FP registers is faster.
> > Neil Roberts review:
> > - include changes on all places that are storing attribute values.
> > - check with and without -O3 compiler flag.
> > Brian Paul review:
> > - use fi_type type instead gl_constant_value type
> > - fix a bunch of nit-picks.
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=82668
> > Signed-off-by: Marius Predut <marius.predut at intel.com>
> Looks good. Thanks.
> Unfortunately, there's a few new compiler warnings after applying the patch.
> I'll try to fix those up before I push the patch. I'll also do some piglit
I am happy to hard that :-). I am looking to see it upstream!
Yes, I see, I forgot some Floats instead file_type especially on _playback_copy_to_current but not only.
I will send a new patch that, additionally, add "fi_type" type instead glfloat where I escaped.
More information about the mesa-dev