R520 - no FP texture filtering?

Originally posted by Humus:
I wrote 3 of the new papers (HDR Texturing, Programming for CrossFire and Framebufer objects).
I haven’t found the papers apart. I use Linux so there is not much use for your SDK but I’m interested in the papers.

regards

marco

Actually, all OpenGL samples have Linux Makefiles and should compile and run fine.

Originally posted by Humus:
Actually, all OpenGL samples have Linux Makefiles and should compile and run fine.
Perhaps ATI should provide it in a more widely, and freely, usable format then, than a Win32 .exe?

.zip (not nice for such amounts of data), .7z or .tar.bz2 comes to mind.

Just an idea…

Is there a list of the OpenGL extensions supported and hardware limits (texture sizes, uniform storage size, shader lengths, etc) for the r520 series anywhere? Google’s not being my friend :slight_smile:

If it emulates FP filtering via shader tricks, it’s fine for me.

I personally liked very much the extreme branching capabilities. SIMD/MIMD branching is bad and this seems to cut it quite well (although I believe a 16pixel batch is possibly sub-optimal but I’m not really expert on the HW side).
I am personally very confident on this generation, although I still like more GeForces. At least they’re now pushing for real.

EDIT: as for VTF, I’m not sure it’s really useful for real world cases yet. I see it worth for stream processing and such but unless you get the array back for collision detection, all this macroscopic detail is going to feel buggy.