[GEM-dev] RGB32

Cyrille Henry ch at chnry.net
Mon May 19 15:26:15 CEST 2014


hello,

according to the mailling list, there is a mails that I did not received.

so, i added a framebuffer type : RGBA32.
The only 32 bits type available in Gem was GL_RGB_FLOAT32_ATI that was introduce in order to make some bump mapping. Latter, 32 bits type was used for physical modelling or other GPGPU stuff. But this type is not available on intel gpu, at least with current driver under linux.
So, i just added a 32bit type that can be used.
i did not tried it for bump mapping (years ago, GL_RGB_FLOAT32_ATI was the only one allowing bump mapping with good performance).

It would be nice that Gem choose what type to use, but I don't know how to do this choice when many type are available. I think they are not equivalent depending on what you want to do. So the easiest solution is to test them on each combination of patch / hardware.

moreover RGBA32 is a 4 channel type, while GL_RGB_FLOAT32_ATI is only 3.

cheers
cyrille

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 2014-05-13 12:58, Cyrille Henry wrote:
>
> add a new 32 type for framebuffer (since GL_RGB_FLOAT32_ATI is not
> usable with intel GPU on linux)

wouldn't it make more sense to have a single float32 type on the
patch-level ("RGB32"), and use whatever is available on the given GPU?

mfgasdr
IOhannes



More information about the GEM-dev mailing list