Are Nvidia Really Better in OpenGL?

[Goksly]

Or do they just enforce more optimisations and make it very difficult for you to turn them off?
http://www.beyond3d.com/forum/showthread.php?t=28798
translated - 3dcenter
http://www.nvnews.net/vbulletin/showthread.php?t=65727

The scary thing about this one? It could very well put into question every single OpenGL benchmark since as far back as the NV30

If someone has an nvidia card and an opengl game I would like to see some results. So much for "high quality" huh.

 

[Goksly]

Graph showing performance hit:

With optimisation (i.e. fake high quality):


Without optimisation (i.e. actual high quality):

 

[Sem]

arnt you a known ATI fan boy

not that i want to flame or anything just wondering

 

[Goksly]

if i was a known fan boy.... wouldnt you already know?
oh I get it... you must be one of those "oh he is posting something negative - he must be 'working' for the other side" narrow minded people.
am i right? im right...

if nvidia brought out a product exactly the same as ati in every way, i'd buy ati. if nvidia were 1p cheaper, i'd go with nvidia. thats the extent of my loyalty. as of now, since ive got a tft, nvidia will have to bring a better (read better, not just faster) product to get my money, because as far as im concerned there isnt a resolution higher than 1280x1024 (welcome to the world of cpu limitation)

 

[speeduk]

arnt you a known ATI fan boy

not that i want to flame or anything just wondering

No hes not a fanboy, hes enthusiastic about ATI hardware which a lot of people are.

I would like to see some official response to this though and maybe tests done from a reputable source.

 

[Subtlesnake]

Or do they just enforce more optimisations and make it very difficult for you to turn them off?

This issue doesn't reflect a change in any of the optimisations, so why bring up the argument?

Both vendors use optimisations, and both vendors allow these to be turned off, it's just that in this case there seems to be a driver bug that prevents this in certain situations.

At the end of the day though, nearly all benchmarks are run at default settings so Nvidia's performance lead is unaffected.

 

[Goksly]

Both vendors use optimisations, and both vendors allow these to be turned off, it's just that in this case there seems to be a driver bug that prevents this in certain situations.

Please READ the threads I posted on my original post. The point of this thread is that 99% of users will think that they have switched off the optimisations but they havent. I dunno about you, but other than benchmarking, I always have my sliders on quality and optimisations off. However, the point of this thread (just restating for you) is that just because the little check box says "off" it doesnt mean theyare.

Thanks for your contribution towards this thread. I cant remember the last time people didnt take a thread, targetted towards a bug in a series of driver sets, personally. Its quite mind boggling / assuming.

 

[Subtlesnake]

I was responding to the title and the first line "or do they just enforce more optimisations". If you're just trying raise awareness of this bug then fine, but you seemed to making a point about optimisations on Nvidia hardware in general.

I was just claryfying that there isn't any fundimental difference between ATI and Nvidia regarding the ability to disable optimisations, and that this is a specific bug (from the thread "an old bug in NVIDIA control panel") that Nvidia simply haven't fixed yet.

 

[Goksly]

Ok but either way its a bug that gives a performance gain at the cost of image quality? And its a bug that has been around since (apparently) Feb 2005. If you think thats acceptable... well enjoy the horizontal lines.
Obviously both sides use optimisations, but having the choice is very important, and this is what this thread is about... also a lot of benchmarks are done with aa/af. Nvidia (as the graph shows) will gain a decent performance boost for this, leading to a lot of bad press for ATIs opengl drivers... hmmm

 

[LoadsaMoney]

They are a bit better again, not by much though, nowt much in it now, the x1k series have the new memory controller which they used to sort out the OpenGL on the x1k range, and when the 5.11's came out with the OpenGL fixed (using the memory controller) the x1800's were faster than Nvidia at OpenGL, but Nvida hadve gone back in front again with more driver releases, but as i said nothing much in it now, ATi has finally sorted OpenGL, and it will only get better when they get used to using the memory controller, as the 5.11's was the first time it was used.

The days of 'ATi are crap in OpenGL' are long gone.

 

[Subtlesnake]

Ok but either way its a bug that gives a performance gain at the cost of image quality? And its a bug that has been around since (apparently) Feb 2005. If you think thats acceptable... well enjoy the horizontal lines.

I'm using an X1800 XT

I would agree though.

Obviously both sides use optimisations, but having the choice is very important, and this is what this thread is about... also a lot of benchmarks are done with aa/af. Nvidia (as the graph shows) will gain a decent performance boost for this, leading to a lot of bad press for ATIs opengl drivers... hmmm

Well, nearly all (reviewers) benchmarks are run at default quality, where both sides are using optimisations. In any case though, AA is where ATI shines in OpenGL games (after the 5.11 drivers) due to the X1800's superior memory controller.

I think the general performance (no AA or AF) is what people have an issue with, and why ATI gets the reputation it does for its OpenGL drivers.

 

[Jimbo Mahoney]

Interesting.

And a useful time for:

 

[ACESHIGH]

Well for top end cards, there are reviews with AA /AF & high quality settings, thats why these cards are purchased for are they not. They are done to show people what to expect from these high priced cards in terms of absolute performance at max quality & usually at a high res. Thus if people buy Nvidia cards based on the assumption they are quicker at these high quality settings (& people paying big bucks want quality not just performance) then they may well be getting the wrong cards.

BTW seems Nvidia havent fixed many different bugs in their drivers for a while now. Whats up with that?

 

[egt]

Just read thread thread and it appears that manually disabling optimizations has little affect on FPS. Nothing worth fussing over then.

(& people paying big bucks want quality not just performance) then they may well be getting the wrong cards.

I agree with this. Maybe you should read this thread concerning ati IQ. Whats up with that?

 

[mmj_uk]

Interesting.

And a useful time for:

 

[LoadsaMoney]

Oh my god its that rage3d link again, i don't even need to click on it to know thats that what it is.

 

[Stringy]

I'd rather have high fps and a slight IQ discrepancy than worse fps and slightly better IQ, I know which one contributes more to the immersive experience

But then again the whole Ati vs nvidia thing is 'serious business' so I guess we need to get to the bottom of this.

 

[Dutch Guy]

Both companies use optimisations nowadays, some will be totally invisible to the user and some can show a difference in IQ.

If I look closely at the two images in the second post I can see a difference but I doubt you would notice it when you are playing the game, if you would not highlight the area and did not say which was which it would be even harder to see the difference.

 

[man_from_uncle]

I agree with Dutch guy. If you'd not highlighted it I'd not have noticed the difference.

 

[Steves_pc]

without trying to start an argument or anything but smack bang in the middle of a shooting frenzy in whatever FPS type game would we really notice the smaller details??

Only mentioning as i dont have a super gfx card and tend to leave the settings on a high res but at a playable frame rate and not sure i'd notice small details like those in the pics from 2nd post, frame rates are surely more important than the detail sometimes even if theyre achieved a bit cunningly.