Nvidia Delivers, Where's Ati?

it will be interesting to see how much anistropic performance increases with these new drivers... especially compared to the 8500's impressive anisotropic performance already there...
 
Yes, for sure. I doubt GF4 aniso will ever really approach the efficiency of ATi's methods though, purely because of the differences in implementation. But if they can bring it within spitting distance then that's pretty good news (for GF4 owners). The aniso bug is a little like the 8500 "high-poly" bug; very mysterious and extremely frustrating, with claims that it has been fixed etc popping up all the time. Quite funny really.. :D

MuFu.
 
Don't be too hopeful, it's going to be impossible for nVidia to get the Aniso much faster because of the way it's implemented. It's like ATi trying to get Smoothvision faster, they could, but not by much.
 
What has NVIDIA delivered? I dont see anything. :)

I have the 29.42 WHQL leaks (the article points at the 28.32 and prior) and my anisotropic performance is still the same on the 4600.

The only thing I've noticed is that 8x anisotropic filtering is nearly useless, and completely useless in Direct3D. The framerate drops a tad, but little to no visual effect occurs in Direct3D.
http://shark_food.tripod.com/nfs/nfs.html

It's already damn hard to see with OpenGL.. and seems to be non-existent, "Placebo" minimal framerate drop but with no visual difference for D3D. Are these 29.42's the big "Inquirer" delivery scheduled for "today" that didn't get delivered?

Ciao!
-Shark
 
Sharkfood said:
What has NVIDIA delivered? I dont see anything. :)

I have the 29.42 WHQL leaks (the article points at the 28.32 and prior) and my anisotropic performance is still the same on the 4600.

The only thing I've noticed is that 8x anisotropic filtering is nearly useless, and completely useless in Direct3D. The framerate drops a tad, but little to no visual effect occurs in Direct3D.
http://shark_food.tripod.com/nfs/nfs.html

It's already damn hard to see with OpenGL.. and seems to be non-existent, "Placebo" minimal framerate drop but with no visual difference for D3D. Are these 29.42's the big "Inquirer" delivery scheduled for "today" that didn't get delivered?

Ciao!
-Shark
Click to expand...

I think there's a huge difference in the Morrowind shots with and without AF. Just look at how much sharper the image is.

As for NFS, try switching to the bumper camera view. There, anisotropic filtering is a must, or else the lane markings get real fuzzy real fast. When driving, you're always looking ahead, and I hate it when it gets all fuzzy. You should notice a very big difference.

Why didn't you include the shots without anisotropic filtering for NFS and Q3?
 
Why didn't you include the shots without anisotropic filtering for NFS and Q3?
Click to expand...

Just a simple graphical depiction.

(The Point)<-------------------------------------------------->(Your interpretation)

:) :)

Seriously, my fault for not being very clear. :)

Here's the poop as we know it:
* Alleged drivers that "fix the anisotropic filtering performance issue"

* 29.42s come out with a substantially "lower" anisotropic filteriong performance hit. (which turns out to be an absolutely stunningly small hit between 4x and 8x only. none->2x->4x *performance hit* is exactly the same it has been since the products inception.)

* There is absolutely NO visual difference between 4x and 8x (in Direct3D).

* There is absolutely LITTLE visual difference between 4x and 8x (in OpenGL).

* The 8500 goes up to 16x anisotropic filtering.. and at this maximum level of anisotropic filtering, it's performance hit is *less* than the hit encountered on the GF3/4 at simply 2x.


So, if you figure max anisotropy carries, say a 27% hit on one, and (before) carries a 89% hit on the other.. the simple thing to do to fudge "better" anisotropic filtering performance is to simply cut the process... which apparently has been done on the GF4. Even in Quake3 where 8x does "something"- it can't hardly be miscontrued as 2x the number of texel samples to the scene. And in Direct3D, a smallish (3-8fps) difference in performance that yields absolutely no change at all.

That make sense? I'll be happy to clear up anything else I may not have been to clear on. :)
 
Dang, not ONCE have I been able to access your site Sharkfood. I always see you posting pics all over the place but I can never see them.
It always loads for ~5 seconds and then I get the "Page cannot be displayed". :(

Hehe I can hook you up with 10 MB if you wanna? No limitations, banners etc. Decent bandwidth. (4 Mbit/s) ftp uploads, ftp/http access.

Let me know if you want it and I'll PM or e-mail it to you (the server info that is).

I don't use it and I would very much like to access your screenshots so just let me know. :)
 
Sharkfood said:
Are these 29.42's the big "Inquirer" delivery scheduled for "today" that didn't get delivered?
Click to expand...


Well, the Inquirer strikes again. Inq described these drivers as a 32.x series I believe though. So no, I do not think they are referring to any drive you already have. And if it is true that they are 32.x, well those big number jumps like that usually do hold some performance advance.

Interesting Sharkfood that you comment on an article you obviously haven't read. I mean, if you had, you would have noticed the last word (or number) of the first sentence.

nVIDIA IS TODAY SET to introduce the driver that we all been waiting for - the New Detonator, 32.
Click to expand...
 
Interesting Sharkfood that you comment on an article you obviously haven't read. I mean, if you had, you would have noticed the last word (or number) of the first sentence.
Click to expand...

I read it fully.

Funny how you assume on a generic posting of "Detonator, 32" somehow means a 32.xx detonator revision. So, in your book the Detonator4's are version 4.xx? tsk tsk. Not the case.

It would make perfectly logical sense that some form of "Detonator 32" would be a lingo term for the next series of improved Dets (probably some reference to "32 bit!" for color or whatnot) rather than a declaration of the series number, especially since 29.XX/30.XX/31.XX havent been released nor designed yet.

Cheers,
-Shark
 
Sharkfood said:
especially since 29.XX/30.XX/31.XX havent been released nor designed yet.
Click to expand...

Wrong, 29.xx have been leaked already, and 30.xx and 31.xx have probably been made but not leaked yet or released.
 
Sharkfood said:


Which part of "released" did you not understand? www.nvidia.com still has the 28.32's as the most current released drivers.

I happen to be running the 29.42 WHQL "leaks"
Click to expand...


Hey, you said "released nor designed for that matter"...which is a little bit different ain't it? How can you say a driver set isn't yet designed if you're running it on your box? Regardless of it's been released or leaked?

Unless you meant to use the word "designated" instead of "designed", in which case it makes perfect sense and I'm just being a grammer-nasty.

(Forgive me Shark, I'm dooming over the lack of a detonator 5 release.)
 
Well, my previous post here was pretty clear I'm using the new "leaked" 29.42's.

I too am bummin' we dont have some new drivers that actually DO something. hehe. Not to mention the 28.32's are getting a bit crusty on NVIDIA's site. The GF3 saw LOTS of driver drops in the first 2 months- just kinda wondering why the GF4 isn't getting the same attention.

They are probably too busy trying to bribe John Carmack with more learjets and ferrari's to say something profound about the upcoming NV30. :)
 
You must log in or register to reply here.
Back
Top