^eMpTy^ said:Hate_Bot...seriously...if you want to have an ATi-love-fest thread...just put it in the ATi section and you won't get people raining on your little imaginary parade...
Actually they ruin threads in there too
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
^eMpTy^ said:Hate_Bot...seriously...if you want to have an ATi-love-fest thread...just put it in the ATi section and you won't get people raining on your little imaginary parade...
I know Ubi did a good job with the SM1.1, it's like the HL2 SM2 or not screens- brighter light, shinier pipes.{NG}Fidel said:Sorry about that Rollo your right about HDR, still that tech is not as optimized as I would have hoped. Empty is right too though, HDR can still be done on 24bit even if it does not look as good. The Diffrence is not really noticeable unless you compare screen shots from in game. But ingame is a diffrent story, you wont notice one thing, but you will have AA and AF at a High Resolution.
The thing is Ive compared in game screen shots from both pixel shader versions and while 3.0 looks better I cant say its leaps and bounds better. Dont let the PS1.1 moniker or whatever it is fool you, they did a good PS2.0 quality job with it.
Unfortunately, that is not what Crossfire does at all.Imagine the SLI profile creater created game profiles for every game you played with out you lifting a finger? That is no profiles, and like it or not it is much much simpler than having to make your own profile.
After talking to ATI's Catalyst driver team, we learned that although CrossFire's Direct3D and OpenGL defaults allow it to accelerate any application, the drivers will also ship with application-specific profiles. These profiles generally use Alternate Frame Rendering, which apparently offers better performance than superTiling or scissor modes. However, AFR apparently can't be blindly enabled without the danger of causing display corruption or stability problems, so scissor and supertiling modes provide good fallback positions for games that lack AFR profiles.
The thing is, they can't say it will accelerate any application. If an application is cpu limited with one card, it won't be running faster with more GPU power?
2Fresh said:Is it that hard for you guys to have a thread that is on fucking topoic???? Ignore the troll and move on..very simple. Why in the hell do you guys think threads are always locked???...for stupid ass reasons like what is going on in this thread. He is a troll, we all know this. MOVE ON>
Terra said:3. How can something with no release date pick up "momentum"?
Terra...
^eMpTy^ said:Nah...we kinda tag-team it really...Rollo takes MWF and I get TT and we both do whatever on the weekends...
Why do you ask? Did we miss one?
madmat said:It's (FP 32) 32bits per calculation, FP24 is 24bits per calculation and FP16 is 16bits per calculation...it's a reference to the level of accuracy of the shader units.
In the shader language that ATI employs the calculations are done quicker and with less accuracy than nV employs because nV employs full 32bit precision which means the GPU is doing more work per pixel in rendering.
Personally I think nV should have opted to allow 12, 16, 24 and 32 bit precision to allow the hardware to take full advantage of somewhat sloppier code, but hey, that's just me.
As to my earlier post I'm just calling it as I see it man, you weren't saying that FP32 was 128bit, you said several times it was 12bit which by your math equals, what? 3 bits times 4?
^eMpTy^ said:I'm looking at my posts and I don't see 12bit anywhere...
What I said was correct but I was taking it somewhat out of context due to my own shaky understanding. And I think NV does support FP12 or FX12 or whatever it is...