View Full Version : Unwinder not to offer driver level support for Radeons
Babel-17
01-09-03, 04:10 AM
but he will offer some tweaks.
Unwinder
Moderator
Registered: Sep 2000
Location:
Posts: 1096
Re: Radeons and RivaTuner
quote: Originally posted by Milenko
Will tweaking of radeon cards be implemented into rivatuner? Or will it strictly be left ot OCing radeons?
Only low level ATI tweaking (overclocking, refresh overrider, diagnostic and gamma correction) is (and will be) available in RivaTuner. Next release of RT will also contain patches for ATI drivers.
Unfortunately I just have no time for supporting driver level ATI tweaking. Correct driver level tweaking requires disassembling each revision of driver and full understanding of all the registry entries used by the driver.
__________________
Alexey Nicolaychuk aka Unwinder, RivaTuner creator
Randell
01-09-03, 04:25 AM
well as fast AA and AF are not hidden settings with the 9700Pro - I'm not that fussed I dont use a tweaker now and dont see the need to use one with the 9700Pro.
Babel-17
01-09-03, 04:51 AM
True that. :)
Unwinder is a magician though, lol, sort of. :)
I mean ..... first giving the quadro software hack to nVidia users (as well as major aniso tweaks and extra aa options etc.) and now enabling "hidden" pipelines on the R9500 ..... you bet I would have liked him breaking down the drivers for us.
There are good reasons for ATI and nVidia not giving the general public a plethora of options, it's all to easy for the public to fubar their systems as it is. :) Lol, but since when are slackers the general public? ;) :)
More seriously, it's not inconceivable Unwinder would (for example) work out more optomized aniso routines ..... not saying anything he would do would really be all THAT but he does look at drivers from a hardcore perspective ..... that being, more options are good.
I just had a thought. :) Yeah it almost died of lonliness ;) but what if Unwinder found a way to enable hybrid ss/ms aa? What if it was the "holy grail"* of rgss aa?
That would be a major boon for older games like flight sims or games that are hideously cpu bound (lol, like some newer flight sims). :)
*jk
Randell
01-09-03, 04:54 AM
Ok if Unwinder or similar could enable a PJGMS/PJGSS mode on teh 9700pro - I would use that tweaker a lot :)
And yes software enabling a 9500 into a 9500pro is amazing if it works as advertised.
Babel-17
01-09-03, 04:59 AM
Lol, yeah a lot of 3dfx refugees would, I suspect.
Babel-17
01-09-03, 05:04 AM
It might not be that more amazing than that hyper-z tweak from the original Radeon.
Btw, I don't really know that much about this stuff, I just surf the boards. I've heard the word "decompilers" bandied about at times. Is that something Unwinder might be using? Can you (or anyone) give me some insight on that? Are we talking mountains of work and lots of inspiration?
Feel free to dumb down the explanation a LOT. :)
Babel-17
01-09-03, 05:37 AM
Hmmm, farfetched, but a remarkable coincidence is not entirely out of the question. ;)
Lol, speak geek?
http://www.rage3d.com/board/showthread.php?threadid=33658612
Ailuros
01-09-03, 07:27 AM
Is it just me or has Rivatuner RC12 been stuck w/o an update for the longest time ever?
Randell
01-09-03, 07:28 AM
I understand the latest beta's included 16bit aa so it could be that.
enabling a MS/SS hybrid would be a great bonus just now as the one AA positive the Fx has over the 9700Pro are the xS modes.
If ATi took away that advantage - a bit like the Det 4's spoiled the 8500 launch eh?
One can hope - I understood things like 16bit AA/SSAA were lower priority compared to game fixes, but it appears that they are being worked on now as there are fewer game issues left.
Babel-17
01-10-03, 02:02 AM
I'm wrong again. Oh well.
[Zardon] Will The Catalysts for the 9700 support both multisampling and supersampling anti-aliasing in the near future, such as detailed here http://mirror.ati.com/vortal/r300/educational/main.html?
[Ben] We have investigated this possibility, however we believe that sample-for-sample, the RADEON 9700 & 9500 series multisampling implementation, provides better performance and image quality than any supersampling method. So we currently have no plans to support supersample anti-aliasing on these products.
http://www.driverheaven.net/ben
Randell
01-10-03, 02:05 AM
goddamn thats crap - anyone knows thats crap.
Babel-17
01-10-03, 05:08 AM
Yeah. Well, maybe we'll eventually see it as an unsupported feature. I'm enjoying the positives too much to be really bummed out. I would have much preferred it though had he dealt with this head on and delivered an apology, no matter how oblique.
"Sorry for giving rise to false expectations, it's just not in the cards ....." for example would have been enough. ATI should realize this does have a bearing on how we the public will receive further such pronouncements on "promised" features with cards in the future.
Lol, will they have to tack on a "really this time, we mean it" at the bottom of every page? ;) :)
Randell
01-10-03, 07:17 AM
hmm what I disagree with si teh smple for sample comment about better image quality. Yes its a matter of degree but an SS solution would help in quite a few cases.
Babel-17
01-10-03, 07:47 AM
Yes, and the lod bias could be pushed a notch also. I don't consider the loss of plain ol' ogss a really major loss, it was the thought of jittered rgss or some other kind of advanced ssaa that had me excited.
Well, on to better things. :)
Ailuros
01-10-03, 07:52 AM
Randell,
Realistically in latest drivers the 8k5's SSAA resembled more to OGSS with 2x samples than anything else. Why the difference in terms of EER from early to recent drivers was/is so drastic will most likely remain an unsolved mystery. It doesn't make sense why it happened.
Of course could they work on a hybrid MS/SS sollution, by just adding garden variety SW based 2xOGSS to the mix, but Supersampling has a fillrate impact (which MSAA hasn't) and cannot make use of any colour compression. 4x mixed mode would then turn out significantly slower and I'm not so sure that ATI is willing to leave NV much headroom to claim too many advantages if any.
IMO it's safer to leave NV brag just about HRAA/2xRGMS performance. I know it sucks for the end user. but they are determined to be as "ruthless" as possible.
edit: and to avoid any misunderstandings; Supersampling creates fillrate concerns above all; my lucky guestimate is based on the fact that even at 400MHz the NV30's fillrate is higher.
Randell
01-10-03, 09:44 AM
Ail,
yeah maybe the fillrate concerns are real enough - but saying no SS solution provides better IQ than their current MS solution is just plain wrong.
Oh well - its amazing how easily you can learn to ignore alpha textures..
SirPauly
01-10-03, 09:56 AM
How about an apology for touting a quality super-sampled AA but instead we get this: "we have investigated this possibility."
So we currently have no plans to support supersample anti-aliasing on these products.
At least the future we may see some innovative super-sampled as fill-rate and bandwith numbers increase.
It is pretty tough to be bummed out like Babel is offering because the product truly does deliver on many fronts. A super-sampled option would be the ultimate icing on the 3d cake for me for certain titles.
Babel,
I'm enjoying the positives too much to be really bummed out.
Remember our discussion on super-sampled and I was kinda torn somewhat? The many positives of the product placed super-sampled on the back burner for me over-all. I wanted it but the combination of a real high resolution with high quality levels of AA and filtering was just a leap for me compared to the past and more than fine for now.
Simply put: I 'm enjoying the positives too much to be really bummed out, hehehe.:)
SirPauly
01-10-03, 09:59 AM
What if the super-sampled on the 9700 is fubared like the 8500's super-sampled?
And ATI doesn't feel like spending the time and resources on work arounds considering the multi-sampling works great?
edit: The reason I am thinking this way is because ATI is sounding really PR with their answers about super-sampling.
Ailuros
01-10-03, 10:16 AM
At least the future we may see some innovative super-sampled as fill-rate and bandwith numbers increase.
Pauly,
If game complexity would remain idle at the point where it is today, chance are higher. But with all the shader wizzbang that's being introduced supersampling could be a tougher cookie for accelerators than ever before even if they sport 40GB/s bandwidth. Most of it will be wasted on multiple render targets and floating point precision anyway.
What if the super-sampled on the 9700 is fubared like the 8500's super-sampled?
Uhmm can I make an assumption in the other direction? What if R300 never gets Supersampling for the same reason R200 never got Multisampling? ;)
SirPauly
01-10-03, 10:36 AM
Well, I hate to think this way.....but when you get idiotic responses from bright individuals, well, they're usually covering something up.
Ailuros
01-10-03, 11:08 AM
One cannot have it all. All of you experience on a daily basis that the R300 delivers more than it actually cost you. I realize that the urge for SS is legitimate, especially since it's been mentioned in an official presentation.
Most of you have your old cards I figure; wouldn't building a secondary system for those older games be an option there? I'm seriously considering it for this year to do something similar. I always just sell or give old hardware away; most of the times it's not even worth the money selling it.
Babel-17
01-10-03, 11:10 AM
Ok, good rational posts, me like. :)
It's tough for me to read between the lines because there are so few lines to read, period. Lol, and I don't have the tech savvy so I'm at a double disadvantage. It's always to easy to read what one wants into a situation, especially when it's an emotional one.
This ATI dude doesn't waste any syllables if the topic is controversial. I can not with any justification claim to have gleaned any crumbs of insight. Lol, very tight lipped and he really ruled nothing out, this is smart as it leaves no firm handle for any antagonism to grab on to.
There can be no "Ah ha!"'s when he doesn't say anything. No "we're not supporting feature "X" because" ...... a) "time is money and not many people are jonesing for it" or b)"this is a tough one, we're stumped, one minute it works ...." or c) "lol, did we really promise that? We must have been eating 'shrooms that day." or d)"the fan-boy sites will twist the numbers and bash us when compared to nVidia." or even e)"yeah, we suck, we never had any real intention of implimenting it, it just looked so cool up there on our website though ..... I can remember when people had a sense of humor ..... fuggit, I'm outa here." :) :)
Lol, in short, anything people can react to. Instead he's a basilisk and it's like listening to someone in the hotseat at a senate inquiry. :)
I give the interviewer credit for making sure he got the question out there. ATI will be hearing it again and gamers don't like hearing legalese. Better to just 'fess up imo and say almost anything ..... "performance stinks and tons of errors" Ideally they would say, "in that one area we bit off more than we can chew, we've learned a lesson here."
In all honesty I must say they delivered more than I expected and the driver support appears good so far, maybe even very good considering. But like many here I get my back up a bit when the normal marketing hyperbole we've all grown to know and love ;) graduates to instances of double speak. Let's (ATI and its customers) just acknowledge the situation for what it is and move on.
I do wonder just a teensy bit if the tight lipped responses aren't due to a little bit of a legal grey area in what their obligations are. I mean no more than just that, I'm happy but this is one litigious country hehehe. :)
dexter96
01-10-03, 03:16 PM
Most of you have your old cards I figure; wouldn't building a secondary system for those older games be an option there? I'm seriously considering it for this year to do something similar. I always just sell or give old hardware away; most of the times it's not even worth the money selling it.
U have a point here... I am planning to do exactly the same this summer. I have a Voodoo 3 2000 lying around here. It was the most troublefree computer part I had ever owned no matter what drivers I tried on it... the only problem I ever had was exiting Tomb Raider 1... it was solved with the 1.05 drivers I think...
When I replaced it I really missed it but I have the feeling it will be back in action sooner or later...
Randell
01-10-03, 04:08 PM
hmm well I have an older system, my 8500 is in but its only a duron650@900 and remember how crappy the 8500 2xAA edge quality now is.
Babel-17
01-11-03, 04:46 AM
Hehehe, turns out the lumpen proletariat are disatisfied in places other than slackercentral. :)
http://www.rage3d.com/board/showthread.php?s=&threadid=33658992
Ailuros
01-11-03, 05:48 AM
Originally posted by Randell
hmm well I have an older system, my 8500 is in but its only a duron650@900 and remember how crappy the 8500 2xAA edge quality now is.
You're meaning to tell me that you can't find the first series of drivers for the card? The purpose of building such a system would be mostly to play older games, for which you hardly should need driver or other SW updates.
Since we're talking about SSAA, the CPU is of less importance there too.
Randell
01-11-03, 04:19 PM
after I posted that I did think about digginh the drivers out again. Just need to think what I've got on the system - OGL had quite a bit of perfromance increase with later drivers.
hmm might pplay aorund a bit, of course the crappy monitor on that system isnt a winner either :(
Ailuros
01-12-03, 12:01 AM
of course the crappy monitor on that system isnt a winner either
Even the better!!! You won't notice the edge quality as much. :D
Randell
01-12-03, 02:31 AM
aye built in AA a feature of blur o vision monitors inc.
vBulletin® v3.8.1, Copyright ©2000-2013, Jelsoft Enterprises Ltd.