BGRA is the most common ordering for both ATI and NVidia parts, however,
you should not see any difference in performance if your components are
unsigned bytes.
Canonically and historically speaking, GL_UNSIGNED_INT_8_8_8_8_REV and
GL_BGRA have been the recommended ordering. Personally, I order things
as GL_RGBA and give a component type of GL_UNSIGNED_BYTE because
it's the easiest way to consider the data arrangement.
Consider also, that internalFormat has to be paired with format/type in
order
to get the best performance. Once the data is in VRAM, the
internalFormat
doesn't make any difference as it is only a request. The driver picks
a format
it likes and attempts to match as closely to the internal format
request as it can.
On upload, the pairing of the format/type with the internalFormat is
important
because it is this pairing that dictates what logical path is traversed
and therefore
the effective throughput.
On Nov 23, 2004, at 1:16 AM, Ciaran Wills wrote:
Hi,
Just a theoretical (or quite real? :) ) question: what internal format
is the best regarding the performance? RGB(A) or BGR(a) ?
It depends; have a look at Nvidia's GPU programming guide, which tells
you what formats are supported natively by which chips.
It's a real question if you're uploading a lot of textures (e.g. video
to texture), as the driver will have to convert the texture to a
native format if it isn't already in one.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Mac-opengl mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/mac-opengl/email@hidden
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Mac-opengl mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/mac-opengl/email@hidden
