Point taken Zalon, maybe because ATI cards are cheaper is why some people buy them? FWIW I get 308 fps stable in qw.. is that not enough? I'll give that d3dquake a try Spike.

EDIT : d3dquake seems to be capped at 500fps. Trying to go higher just seems to force it back to 500. Runs great though Some command I'm missing?

All those posts say it's ATI's fault and that their OpenGL is crap, or ATI cards are crap in general even...
That is exactly the kind of thing we SHOULDN'T be saying, because.... well... we don̈́'t know?
I also find it wrong to tell people "don't buy ATI if they are planning on playing Quakeworld - period". ATI can be fine for Quakeworld (I know I don't have any complaints). What we should be saying, and what I was saying, is that right now, with the current Quakeworld clients and the current ATI drivers, NVIDIA has an edge over ATI in raw Quakeworld performance for similarly specced cards. So if you ONLY care about Quakeworld performance, you get most value out of going NVIDIA. I am sick of statements like "ATI sucks", "ATI's OpenGL sucks", "ATI is crap for Quakeworld" etc.

If you're sick of statements like that, might I suggest that you don't even bother clicking on a forum link on quakeworld.nu thats entitled "ATI openGL problems". Three reasons why it might not do you any good:

1. ATI sucks for QuakeWorld
2. ATI sucks for OpenGL

Several different benchmark tests state this. Sure, you could argue that you can run QW smooth even on an ATI card, but that is after removing most GL-rendered effects from the game. Also a NVIDIA card in the same price range would run QW much better even with all these turned off. You really seem to forget you're in a QW forum?

3. People that play QW, that is also hanging out on a QW related forum, wouldn't mind to point out these two first reasons for a warning / info to other members of that forum, just to let them know that if they plan to play QW, NVIDIA is the best buy for that.

Forums are a place to f. ex express subjective opinions. In this case these negative subjective opinions towards ATI is even widely documented and backed up. Imho, it would be stupid NOT to let people know about the huge flaws in ATIs cards for playing QW. If you get sick of that, go I would go to a forum that is more likely to support your own view if I were you.

If that is what you think, then you should read the posts again - they all more or less just say that ATI is bad for qw...

Here is a couple of reasons why I might want to read a forum thread with that title:

1. I might learn something new. I had not heard about these memory leaks tom posted about before for example.
2. If the thread had been about an ATI owner having trouble with his QW set up, I might have been able to help, being an owner of an ATI card myself.

There are, however, reasons why I shouldn't visit a thread like this:
1. Seeing a posts like "Always ATI, why quake players keep buying that crap is beyond me " might cause my blood pressure to rise, which is bad for my health.


I would like to see those benchmarks. That would be genuinely useful information. My problem is exactly that people just throw around postulates like "ATI sucks for OpenGL" without backing it up. I wish I had a [citation needed] button that I could spam for situations like that.


No one is disputing that NVIDIA cards have an edge over ATI cards in Quakeworld (I hope), which was exactly what I wrote in my post. Just because we are in a Quakeworld forum, doesn't mean people only care about Quakeworld performance. That's probably why people buy ATI cards in the first place (I know that's why I did it) -- they are good value, if you want to play other games than Quakeworld too.


Subjective opinions like "ATIs OpenGL sucks" are not useful or helpful at all. If you want to post something useful, you should cite those sources that allegedly prove ATI has sucky OpenGL support. Your benchmarks, for instance, were very useful information, although they probably didn't surprise many people. Now, every time someone asks what graphics card they should buy if they primarily want to play Quakeworld with e.g. ezQuake, people can cite your benchmark -- no need for useless subjective opinions.

Also, I won't change forums just because people post silly things that bother me. I would much rather show them why their posts are silly, and hopefully make them post something useful the next time.


FTFY

Actually, I find it very useful. Getting input from other players is far from useless to me no matter how subjective their opinions are. And forums are what they are as well. You can of course whine as much as you want that people wont link to sources every time they make a statement, but I really doubt that is gonna happen whenever a person have the need to express something here.

edit: And tbh, if I knew ATI where that poor compared to NVIDIA for OpenGL, I would never have bought it in the first place. I'm pretty sure, that if I had browsed these forums when I bought my computer (january 2008, wasent registered here until march 2008), I would have went for NVIDIA instead. I think that might be the case for many other QW players buying ATI as well; They're simply not aware how poor they are compared to NVIDIA when it comes to QW performance.

Do not mistake qw.nu forum as something like Wikipedia. Of course all the subjective opinions are by people using different products, they can't really be "proven" otherwise than benchmarking.

During the past 10 years I've owned both ATI and NVIDIA cards, and during that time I have always benchmarked my gfx cards a lot (unlike these days, not too interested in that anymore, I'm too old for that shit) and noticed that there are very few cases where ATI's OpenGL ICD is faster than on-par NVIDIA hardware. Actually, in the past gf4 titanium cards were faster than R9700/9800 cards in OpenGL (QW, Q3, Jedi Knight, etc) even if R9700/9800 were almost twice as fast otherwise. Oh, and it never got better, quite the contrary as performance was decreased from catalyst release to another (Catalyst 3.0 had the fastest OpenGL ICD for those games I mentioned back then).

Back in the past, when NVIDIA was founded, a lot of OpenGL engineers joined NVIDIA (can't remember exactly from where they were, but it might have been SGI) and NVIDIA has always been working closely with OpenGL development, even these days. Just go check NVIDIA's OpenGL site, compare the extensions and OpenGL support to that of ATI's.

NVIDIA's OpenGL support is much more complete and working, however you look at it. I don't see a lot of people complaining about NVIDIA's performance on ezQuake for example, but a lot of people seem to complain about ATI's performance, shouldn't that be kind of a clue already? Rikoll's examples prove the difference quite nicely, and there is very little you can say could change the fact his benchmarks show.

And finally, as for the OpenGL support, why not go file a bug report at AMD/ATI's catalyst site? Last time I checked it was possible, but not a single time I reported a bug for them, was it fixed, so I wouldn't really expect much.

Of course the [citation needed] was a play on Wikipedia. I am not seriously suggesting that people need to be quite as rigorous when writing forum posts here as writing Wikipedia articles. What I *am* stating, is that postulates like "ATI's OpenGL support sucks" is not helpful. If it's not generally accepted, you need to back it up with something, at least. Anecdotal evidence like yours would never work on Wikipedia or in academia, but it is about the best you can expect in a forum in lieu of benchmarks. In other words, I have no problems with that. I think it's good to hear some anecdotal evidence.

But these statements ("ATI's OpenGL support sucks" get thrown around like some kind of truism, even though no one ever seems to back them up with some real evidence, and the people that are the fastest to post these comments seem to be people who DON'T own ATI cards -- that is a bit worrying.


Support for OpenGL extensions isn't as important as supporting the actual standard, which both companies seem to do quite well. Also, looking at that list I see mostly extensions that were useful in the days of the fixed-function pipeline, so not that relevant for new games anymore (haven't bothered to look at ATI's page). There is also a huuuge number of NVIDIA's proprietary extensions related to among other things their Cg language. Cannot really fault ATI for not supporting those, can you?


AFAIK OpenGL implementations have to pass a conformance tests to be able to call them self .. well .. OpenGL implementations. As far as I am aware, both ATI and NVIDIA have full OpenGL 4.0 conformance, so I wouldn't say ATI's OpenGL implementation isn't complete... Maybe it's buggy (I've heard that was a problem in the past -- not sure about today), but at least it's complete.


We've been over this before, I think. I only ever submitted one bug report to ATI, about an OpenGL game called Knights of the Old Republic, and that was fixed in the next driver release. YMMV.


"Don't buy ATI crap!!! Their logo is ugly shit!" -- That is useful to you? Extreme example of a subjective opinion, I know, but I hope we can agree that it's not exactly helpful advice when people want help with buying a graphics card. Don't you think people prefer rational arguments?


How about they try it once? I cannot remember a single time someone in here exclaiming "ATI's OpenGL sucks" backed it up with any evidence at all. And no, just because Quakeworld runs less optimal on ATI's OpenGL implementation DOESN'T mean ATI has shitty OpenGL support in general. Could as well be that Quakeworld is shitty old as uses OpenGL in a way that is brain-dead by today's standards.... or it might be both! We just don't know unless someone back up their statements, just for once.


I am not advocating that people should be barred from advicing people to go with NVIDIA instead of ATI if they want to primarily play Quakeworld. That is exactly what I did in my original post, now wasn't it? I even said your benchmarks were useful! Also that statement that Quakeworld (with the clients we use) runs better on NVIDIA than ATI is generally accepted here, I think. It's even been backup. That ATI's OpenGL implementation sucks in general, I don't think is generally accepted at all.

Well, me personally don't bother much about logos, specially not on hardware inside a case, but who knows? Someone who really cared about logos could then look it up and judge himself. If that is something he want to base an investment on, isn't that fine? I'm personally more interested in peoples opinion about a cards performance in QW, and even a "ATI sucks for QW" could wake my interest, and make me aware of the card being troublesome for many people (since many people seem to write this on different threads here). I think it is nice that you found my little test useful, to others a simple "ATI sucks for QW" statement is more useful. If they're interested in knowing more, they could look up more information themselves. Who are you to judge whats useful information or not, or what is peoples general preference? Can you back that up, that your definition of useful information is the correct one? Being a bit sarcastic here, but I guess you see where I'm going with this. How YOU, personally, want others to write on a forum, is (I hate to break this to you), totally irrelevant for most people here. Use and enjoy the forums for what they are, instead.

Is it so hard to get raz0's point?

It's like people writing complaints and negative feedback about products only because they experienced a long delivery time from their retailer.

Now this forum aint that big, but creating a forum post on a forum with 10k+ readers a day, warning people about buying Razer Deathadder cause it's defect (only to turn out later that the guy forgot to plug the USB-cable in, or remove some protective plastic from the sensor), is kind of uneccessary. In my experience, it's the lack of knowledge from the people behind the wheels that is the problem in many cases.

That was in general. Regarding this matter, it's known that ATI performs worse than nVIDIA in ezQuake atleast. Why that is the case is another discussion?

It would be sad if they both didn't fully support OpenGL 4.0, I'd guess they would be shooting themselves in the leg on professional card market if they didn't fully support the spec. By completeness I meant overall extensions for within the OpenGL, not like NVIDIA doesn't support few ATI_* extensions, or ATI not supporting certain NV_* extensions.

As for the ATI's OpenGL implementation and suckyness, well, dunno. I'd rather say ATI's OpenGL is just worse, more buggy and slower (considerably depending on the software) than NVIDIA's. Not that it really matters anymore when it comes to gaming, since there will be very few OpenGL games in the future, as for the professional segment, well, that doesn't really concern us gamers.

The problem is if people only actually write the first part, but think the second. Okay, so you might say they are being malicious then, but you could come up with examples of it's not malicious intent but just ignorance, and then we have a real problem. Then people start repeating it... fog had a good example of that, if you don't understand what I mean.


That might as well be a one person having trouble and N other people just repeating what he said. It is really dangerous to assume that just because something is being repeated a lot, it must be (or is most likely) true. Think urban myths -- how many of those are true? Sure, seeing "ATI sucks for QW" will make you aware of a potential problem, but it IS useless if you cannot verify it easily. You cannot simply 'look up more information' if it's not readily available.


I seriously doubt that. Since there is absolutely no quantification of the alleged "suckiness", benchmarks will ALWAYS be more useful. It might have been a little useful if it said "ATI is about half the performance of NVIDIA"... but you'd probably still want some kind of assurance that a statement like that hasn't been made up before you make any buying decisions based on that.


Again, might not be readily available...


I think it is in the best interest of these forums that the spread of misinformation and baseless postulates are kept at a minimum. As such I will probably keep pointing out such things (or as you'd probably say, whine about it).

Not sure what you mean. As far as I know extensions aren't really that important since developers don't tend to use extensions that are not widely supported, and certainly not extensions that are not supported by one of the major graphic cards vendors. Extensions have also become *much* less relevant after GLSL was introduced.


Yeah, it's too bad about not many new games using OpenGL. Still I haven't seen any real hard evidence that ATI is so much worse with regards to OpenGL in general. Please do fill me in if you have some links or whatever -- I'd be much interested to read it.

Yes, I know you think that way, but other people just think different, I'd even dare to say the majority. For a review, or a comparision test between ATI and NVIDIA, I would of course agree with you. I get your point, totally, you don't have to explain it over and over. It is just that I think your point is pointless. I'm also sure many would disagree with me and agree with you, if that gives you some peace. This is a forum, and I'm sure many would stop writing at it at all if you're gonna use your magnifying glass on every statement that is written here. I'm also pretty sure it would be extremly tiresome for you over a longer period of time; you've already stated you're sick of it. This if of course all speculations, but hey, we're on a forum, right?

The problem is, no hardware site benchmarks anything using OpenGL anymore, so the results are from the past. However, you can always do some googling by yourself, not that it's going to give any good results.

However from what I've seen, ezQuake has an issue with ATI hardware. Not only that, since I'm an astronomeur and use a software called Stellarium (that uses OpenGL), which again has serious problems with ATI hardware. And like I already said, even one generation older NVIDIA hardware were faster than 9800-cards for example in a number of older OpenGL games (Q3, QW, Jedi Academy, UT99 using OpenGL renderer...). Some specific games could be faster on ATI (especially the latest OpenGL games) that have been highly optimized, like Prey.

For benchmarking these days you would need something like Furmark, maybe Unigine with OpenGL4 support and similarly performing hardware like GTX 470 and HD 5870. Also you could run some older OpenGL benchmarks.

As for the "ATI OpenGL problems", it's clear that people here at qw.nu are having (performance related) problems with dynamic lights on their ATI cards, making ezQuake perform like it's run on Voodoo3 when those are enabled.

How many more benchmarks do you need? Multiple long time qw players have gone through the trouble of testing framerates with both nvidia and ati cards, using the same computer otherwise (I suppose).

I understand what you mean in a broader scope, but here we are simply discussing ATI + opengl qw = suckage, especially with dynamic lighting. Although I trust Renzo's benchmarking quite a lot, I can't speak for his benchmarks concerning other older opengl games as I haven't had a chance to do that kind of comparisons.

Guys have you ever tried to use -nomtex parameter to disable multitexturing on ati card.
on my ATI 5770 it helps a lot. on intel cards too. Big thanks to mhquake blog.
(quest for fast lightmaps series

Hey, the ATI admin is getting fed up with the OpenGL problems but at least admits it's a problem indeed. The last post:

"The OpenGL issue will be fixed soon as I have already said. Thread closed."

http://forums.amd.com/game/messageview.cfm?catid=260&threadid=144563&highlight_key=y&keyword1=opengl


The other topic from 2009 (1st post):

"All bugs submitted are read as I have seen the reports. The more people who report a bug, the higher priority it gets to be fixed. If nobody or very few reports it then it wont get fixed as its not important enough."

http://forums.amd.com/game/messageview.cfm?catid=279&threadid=108194&STARTPAGE=2&FTVAR_FORUMVIEWTMP=Linear



Seriously

As punishment perhaps every single Quaker should open a topic asking what 'soon' actually means in the world of Radeon cards

Spyre is the most retarded fanboy ever (fan@ic).

Back in the day, when doom3 was released, people started reporting about a problem with current 9800pro and 9800xt cards: In doom3, there was white pixel snow all over the screen on a lot of those cards (graphical errors). People took photos of it, and some even took a video of it. At that time I had 9800xt, which also got that white pixel snow.

It appears that a lot of those cards couldn't take the load doom3 caused and artifacted because of it, so the damn hardware was buggy. So, what did Spyre do? He closed the thread and banned the users reporting of the problem because it was:

1) just a flamewar attempt
2) there was no proof whatsoever that the cards were faulty
3) problem with the users and overclocking

The best part was the fact that it happened even when underclocked the cards, most ran at their stock clocks. No official statement was given about the problem, also there was no solution for it either. You see, any problem goes away as long as you kill the messenger and remove all content written about it.

yeh, the kill the messenger part has been tried quite often

Thanks for reporting that, Renzo. Well.... the problem didn't go away this time But I'm sure most ppl are still not aware of the problems with OpenGL games and only read those review benchmarks which doesn't mention those games anymore. But ... I would think that back in 2005 they should have noticed fps drops in Doom3? I might check out some old reviews...

Anyway, it's also sad to say how some Radeon users on that forum seem to think that it's completely normal that you can't properly play old games on new hardware anymore. This was actually also the feedback from ATI given to a user concerning Doom3. They just blamed the old game.

I don't remember there being a performance problem, only the fact that factory clocked cards were clocked too high and they couldn't handle it in doom3, but started producing artifacts.

You shouldn't really read individual users' views of how things should work or not (a game being old and all that shit), but rather try to get in contact with the manufacturer (in this case, ATI/AMD) and complain/file a bug report. If they tell you "it's too old of a game, priority -1" then you know they just don't care.

One thing I've noticed during the years I've been into PC-hardware is that there seems to be some sort of grudge going on if someone is an AMD/ATI user, and they try to attack NVIDIA users constantly. It's almost the same on the CPU front, but not as much pronounced. I'm first to admit that during the past what, 5 years or so, when I have owned NVIDIA cards, there have been some problems with the drivers, but nothing really major in my case. Certain AMD/ATI users never acknowledge any sort of problem with their drivers, even if it was bitchslapping them to the face. Just check any random forum about NVIDIA Fermi cards and you'll see what it is.

Ah, a bit offtopic but anyway...

aye, its a good job game tournaments don't require only ATI cards.

I've found bugs in ati drivers, I've found bugs in nvidia drivers in the same sort of place.
Windows will detect frozen gpus and restart them if the drivers won't, so drivers don't even need to be stable any more.
Both companies were slow to provide vista drivers, and this is one of the more significant areas where vista got a bad reputation, but both sets of non-cheapo-intel cards constantly reset, so it must have been vista rather than the drivers.

Both companies suck.


Even if you can find benchmarks for modern opengl games, the featureset used by those opengl games have nothing in common with most quake engines.

ezquake's dynamic lighting is slower on nvidia and especially intel cards than it should be, too. Its not just ATI. Its just that ATI doesn't quite optimise for compatibility paths so well. mh created a nice informative topic over on inside3d regarding this, along with benchmarks. glquake did it the slow way, and ezquake doesn't replace that, but don't dispair as q2 also does it wrong.
Its actually kinda annoying. Lightmap updates are faster in FTE than q3-style dlights, on ati hardware, or I would have removed them and merged dlights and shadowing.

But yeah, lets all complain about ATI.

The thing is, when was the last time you saw any competitive product from any manufacturer other than NVIDIA or AMD/ATI?

3DFX was one of the top dogs back in the late 90's. Matrox was ok from time to time, but with Parhelia they failed miserably. Sure, it had the specs but it still was slow. There also was a tiler card called Kyro2, which used a bit different rendering method (tiling) and was fast, but for some reason even it didn't manage to stay competitive.

I don't remember a single Intel card that has been in any way competitive with NVIDIA, ATI or late 3DFX. For gaming, they just suck and that's that. I've also read some recent articles about Intel optimizing their drivers for better performance, but at the cost of image quality (a big joke among the scene that follows gaming hardware like GPUs). If Intel's drivers suck for some game, almost noone really cares. So, if NVIDIA can run some older game fine, one should expect similarly performing AMD/ATI cards to do the same game as good. If not, it's kinda easy to blame them for not supporting "everything" properly. Also note that AMD/ATI's OpenGL performance problems are not limited to games, but in other software that might use OpenGL too (for example Stellarium comes into my mind first, since it's related to my hobby).

Well, ATI = now AMD and I like AMD (for everything except GPU), so it's not that I want to bash them for no reason.

[edit]rest was bit off topic [/edit]