nvidia has officially released their ‘long awaited’, cheats free 
50 series drivers but there is nothing about GLSL in them 
I hoped they would add non-rectangle float textures sopport… what are they waiting for? It could be implemented using rect textures at driver side if they have problems with memory manager. About GLSL - I feel a bit uneasy about the specs, I don’t think they are free of bugs till now.
what are they waiting for?
Their hardware.
It could be implemented using rect textures at driver side if they have problems with memory manager.
Not if you want mipmaps.
… or want wrap mode.
Regarding GLSL, it will take some time for it to appear. It’s not a quick and easy thing to implement.
Long Live Cg!!!
-SirKnight
I think, NV implement a GLSlang like CG - additional compiler/run-time.
Also, in this driver exist emulation of NV40 (NV_vp3/NV_fp2), but it doesn’t work :-(.
Originally posted by SirKnight:
Long Live Cg!!!
-SirKnight[/b]
Tell that to someone with a Radeon 8500 .
cu
Tom
Hmm, vertex_program3…I wonder what kind of goodies they added in there. And fragment_program2…oooohhhhhhh niiiice. I’m betting fp2 will be like vp2, if not better, but for fragments of course. Too bad I won’t get to use these in hardware on my FX. But it’s understandable, it’s NV40 stuff not NV30 stuff. 
-SirKnight
Well you could always use D3D with Cg on a 8500.
-SirKnight
Originally posted by SirKnight:
[b]Well you could always use D3D with Cg on a 8500.
-SirKnight[/b]
Of course if you wanted to go to the dark side you could also use HLSL…
Originally posted by SirKnight:
[b]Hmm, vertex_program3…I wonder what kind of goodies they added in there. And fragment_program2…oooohhhhhhh niiiice.
-SirKnight[/b]
oh yeah… i really “like” that they even have to create a versioning/extension sheme to their own extensions…
Korval, i don’t want mip-mapping. If I’m not mistaken floating-point textures support no mipmapping either. Wrap mode could be easily emulated. All I want are floating-point cubemaps on FX! With fp rects and fp cubemaps + ARB_superbuffers - I will be completely happy.
Well well well!
GL_ARB_shading_language_100
Is this in Catalyst 3.9?
And if you tried it, give feedback.
Originally posted by Tom78:
[quote]Originally posted by SirKnight:
Long Live Cg!!!
-SirKnight[/b]
[b] Tell that to someone with a Radeon 8500 .
cu
Tom[/b][/QUOTE]
That’s why it sucks.
Originally posted by davepermen:
oh yeah… i really “like” that they even have to create a versioning/extension sheme to their own extensions…
What the hell are you on about? There’s nothing wrong with doing this. They create a more powerfull version for their cards…so what else are they going to call it? NV_something_new_and_cool ? Sometimes your extreme bias towards ATI is rediculus. Sheesh.
-SirKnight
fishman : i dont think youre going to see glsl on 8500 either…
Originally posted by SirKnight:
[b] What the hell are you on about? There’s nothing wrong with doing this. They create a more powerfull version for their cards…so what else are they going to call it? NV_something_new_and_cool ? Sometimes your extreme bias towards ATI is rediculus. Sheesh.
-SirKnight[/b]
same to you… Sheesh.
i’m just sick of that “woah, nvidia has some new leaked driver with some new leaked version of some new emulator of some new once-useable feature of some new not-existing gpu that we can now access!”. this is so ridiculous.
they should first get their job done well again. there’s no gain in spitting out proprietary extensions for proprietary hw.
rc will die out, and whats with all the nice gf2 effects i’ve coded? nonrunable anymore. i’ve coded stuff with NV_vp, not runable anymore. etc. nvidia proprietary exts aren’t scalable, or really usable. thats why they even need versionings on them.
i don’t believe in that hype. you do. you love them. so continue to love them. love cg, love nvidia. its the way its ment to be played.
i prefer to have it simple, clean, and fast. and future proove.
i prefer to have it simple, clean, and fast. and future proove.
Let’s not forget castraited.
NV_vp2 exposes conditional branches in the vertex program. ARB_vp does not, and yet my 9500Pro could easily do it. It certainly does under D3D 9’s vp2.0. So, why can’t I access these features? Because ATi is too lazy/short-staffed/bound-to-glslang to bother writing an extension to the ARB_vp that provides this functionality.
I suppose nVidia should have just not exposed register combiners, then. Or vertex programs, for that matter. Had they not done so, OpenGL would be in a far worse state than it is now, compared to D3D. For a good year or so, OpenGL’s compariability with D3D as a modern, high-performance API was based primarily on nVidia extensions. Everything from VAR to vertex programs were done on nVidia implementations first.
How long did we have to wait from the launch of the 9700 until we could actually use some of those vaunted fragment programs? 2-3 months or so? Had it been nVidia, we would have had the extension specs a good month before the hardware’s launch, and we would have had emulation in the drivers so that we could already begin coding effects.
I would much rather that they expose hardware-specific functionality to me to allow me the option of using it. That’s something I miss on ATi hardware. Complain about their hardware or drivers all you want, at least nVidia gives you access to all of it.
agreed
(now i want to see the hm?!!? answers )