I owned one for about a year and have been using r300 for 6 months now.
I did propose that I still liked the IQ better on my 8500 then my geforce 4s.
THe 9700 is way beyond that though.
Didn't mean to confuse either of you.
sorry
rogo
I did propose that I still liked the IQ better on my 8500 then my geforce 4s.
THe 9700 is way beyond that though.
Didn't mean to confuse either of you.
sorry
rogo
After reading all this I wonder why the FX5800 was put out in the first place.
If U use the adds to scream Hollywood FX all over the wall, how come that the PS2.0 on the GFX is SO bad
Claudii
"I play games obviously, but it's easier for me to tell which card has more accurate rendering from tests like this than screenshots of games on a website."
Rollo is a strange nut
But an intelligent one.
I don't automatically discount his pov as irrelevant but I do question his sense of "nvidia loyalty" that he seems to hold close to his heart.
Rogo
"I play games obviously, but it's easier for me to tell which card has more accurate rendering from tests like this than screenshots of games on a website."
Rollo is a strange nut
But an intelligent one.
I don't automatically discount his pov as irrelevant but I do question his sense of "nvidia loyalty" that he seems to hold close to his heart.
Rogo
Ok I have only two issues with ben's comments ( I really am not qualified to argue the finer points of AA/fsaa/AF)
Ben says he was told that ATI's drivers were at the level of nvidia's, and then he says he does not belive this to be true. I say that ati has even better drivers than nvidia... in fact I do get more crashes and lockups when testing on nvidia cards than I do on ati cards... I would have to attribute this to the whql testing (which requires strenous testing of drivers before release). If I want to get to the same level of driver with nvidia I have to compare to older tested drivers sometimes changing the scores by quite a large amount. I have used 3dfx, ATI, and Nvidia cards for the last 8 years and buy far the worst drivers that I have ever seen were from ATI in the early 8500 days. I was stupid enough to buy the 3dfx banshee, rage fury maxx, and a matrox mastique and pre-tnt nvidia cards too. The banshee and the pre-tnt and original tnt drivers were not too great than as well.
anyway as far as the texture filturing if you are talking about anisotropic than I would say again to mirror every one that does not have nv-bias, If it does not exist in games than stop crying about it (unless you are using this in your work/hobby and use a custom app that actualy makes it noticeable). This is tantamount to saying that ps 1.4 or truform are of no importance because what games use them. I agree with both of the above comments truform even in the games it is implemented is so hard to tell that it is working because it is easier to just make all the poly's, and as far as PS1.4 it is better again if emplemented fully but why make it just a little bit better for less than 1/4 the cards currently on the market. and as far as AF on more than 3 degrees, why worry when you rarely if ever see this in any game let alone have time to notice it if you do come across it.
I do not have the geforce fx so I can not compare anything to it however I do have the ti's, gf3 and 4's, the radeons all but the 9800, and different other assortments of testing gear. I have to say that I loved my radeon, than my gf3, than my 8500 after the driver nightmare was over, and I loved my ti4400, and I love my 9700pro. the problem is saw reasons to upgrade those cards, weather it be speed or features or both I don't see a reason to go to either the FX or the 9800...
BTW I know this is totally off the current discusion of this thead.. but ever since the radeon there has been no card even qualified to run video to hdtv's the radeon's have just been too dominant here.
well ben and rollo I hope you can find happiness in the nv35, may it beat ati so you can re-find your inner peace... J/K
Ben says he was told that ATI's drivers were at the level of nvidia's, and then he says he does not belive this to be true. I say that ati has even better drivers than nvidia... in fact I do get more crashes and lockups when testing on nvidia cards than I do on ati cards... I would have to attribute this to the whql testing (which requires strenous testing of drivers before release). If I want to get to the same level of driver with nvidia I have to compare to older tested drivers sometimes changing the scores by quite a large amount. I have used 3dfx, ATI, and Nvidia cards for the last 8 years and buy far the worst drivers that I have ever seen were from ATI in the early 8500 days. I was stupid enough to buy the 3dfx banshee, rage fury maxx, and a matrox mastique and pre-tnt nvidia cards too. The banshee and the pre-tnt and original tnt drivers were not too great than as well.
anyway as far as the texture filturing if you are talking about anisotropic than I would say again to mirror every one that does not have nv-bias, If it does not exist in games than stop crying about it (unless you are using this in your work/hobby and use a custom app that actualy makes it noticeable). This is tantamount to saying that ps 1.4 or truform are of no importance because what games use them. I agree with both of the above comments truform even in the games it is implemented is so hard to tell that it is working because it is easier to just make all the poly's, and as far as PS1.4 it is better again if emplemented fully but why make it just a little bit better for less than 1/4 the cards currently on the market. and as far as AF on more than 3 degrees, why worry when you rarely if ever see this in any game let alone have time to notice it if you do come across it.
I do not have the geforce fx so I can not compare anything to it however I do have the ti's, gf3 and 4's, the radeons all but the 9800, and different other assortments of testing gear. I have to say that I loved my radeon, than my gf3, than my 8500 after the driver nightmare was over, and I loved my ti4400, and I love my 9700pro. the problem is saw reasons to upgrade those cards, weather it be speed or features or both I don't see a reason to go to either the FX or the 9800...
BTW I know this is totally off the current discusion of this thead.. but ever since the radeon there has been no card even qualified to run video to hdtv's the radeon's have just been too dominant here.
well ben and rollo I hope you can find happiness in the nv35, may it beat ati so you can re-find your inner peace... J/K
Eagle-
If you could point me out exacting situations I would be more then happy to check that for myself. I can list exacting situations that will take ATi's R300 boards down every time with the latest drivers, and these have already been acknowledged by ATi as being real(not to mention thousands of ATi users have the same issue). The other problems with ATi's drivers are game related in general. Image corruption in certain titles(Q3 engined games, Sacrifice) or significantly sub par performance in, by today's standards, simple games(NOLF, KPC). If you could point me to where I can reproduce these issues I'd be more then happy to try it out. BFG mentioned the issues in SeriousSam and I checked and they were there, then a quick edit to the game cfg file and they were gone. I searched for solutions to the problems I was having for the R300, looked for tweak utilities, emailed ATi, had a lengthy discussion at B3D about it and no solutions were to be had to the problems I mentioned above. If the issues with nV drivers are the type that I have to take 30seconds to edit a configuration file then I'll stick with that.
As far as drivers overall, I had a RagePro card that I picked up back in '98 and the drivers impressed me so much that I didn't pick up another ATi product for five years. The Radeon9500Pro has made a like impression on me as although the drivers are much better then the RagePro was, everyone elses have improved significantly also. That I still have problems in games that are known issues and not isolated to my rig but are in fact commonplace amongst people using the same drivers speaks to where the problem lies. If I had these same issues with nV's boards, you better believe I'd be bashing them about it.
The things is that I can see the errors just fine. I spent years working with 3D viz, I'm a bit sensitive to rendering errors. As far as nV bias, I've taken numerous shots at the NV30's AF too, but noone is getting their panties in a bunch over that
When talking about AF, I am talking about doing it right versus not doing it right. The NV2X boards are the definitive example of doing the actual filtering properly, their only problem is their speed when doing so. The NV3X boards that we have seen to date and the R3X0 boards that we have seen do not do it right.
For that matter, if people really can't see the problems with ATi's AF implementation, why aren't they looking at nV's low quality high performing modes instead of their highest quality settings when comparing the R3X0 boards to NV3X offerings? Because different people see differing levels of issues on different levels. I can see the issues with the R300's AF implementation, although as I have stated previously(numerous times IIRC) their AA implementation is hands down better then nVidia's(again, why noone defending nVidia from my pokes at their issues?).
You show me a board with NV level drivers, NV2X quality AF, great levels of performance and R3X0 quality AA and I will buy it without hesitation(in that order in terms of priorities btw). That board does not exist right now, if S3 came up with it I'd buy it. If ATi came up with it I'd buy it. Hell, if 3dfx rose from the ashes and pulled it off I'd buy that too. My level of priorities has remained fairly constant for many years now, you can check through the archives and see I've been saying the same things(although my big push for hard T&L is now a dead point as everyone is doing it and very well right now) since these board were put in to place(longer then that actually, but these forums go back to '99 to frame where I'm coming from).
You swap my priorities around completely and you have the perfect candidate for a R3X0 based board.
If you could point me out exacting situations I would be more then happy to check that for myself. I can list exacting situations that will take ATi's R300 boards down every time with the latest drivers, and these have already been acknowledged by ATi as being real(not to mention thousands of ATi users have the same issue). The other problems with ATi's drivers are game related in general. Image corruption in certain titles(Q3 engined games, Sacrifice) or significantly sub par performance in, by today's standards, simple games(NOLF, KPC). If you could point me to where I can reproduce these issues I'd be more then happy to try it out. BFG mentioned the issues in SeriousSam and I checked and they were there, then a quick edit to the game cfg file and they were gone. I searched for solutions to the problems I was having for the R300, looked for tweak utilities, emailed ATi, had a lengthy discussion at B3D about it and no solutions were to be had to the problems I mentioned above. If the issues with nV drivers are the type that I have to take 30seconds to edit a configuration file then I'll stick with that.
As far as drivers overall, I had a RagePro card that I picked up back in '98 and the drivers impressed me so much that I didn't pick up another ATi product for five years. The Radeon9500Pro has made a like impression on me as although the drivers are much better then the RagePro was, everyone elses have improved significantly also. That I still have problems in games that are known issues and not isolated to my rig but are in fact commonplace amongst people using the same drivers speaks to where the problem lies. If I had these same issues with nV's boards, you better believe I'd be bashing them about it.
The things is that I can see the errors just fine. I spent years working with 3D viz, I'm a bit sensitive to rendering errors. As far as nV bias, I've taken numerous shots at the NV30's AF too, but noone is getting their panties in a bunch over that
When talking about AF, I am talking about doing it right versus not doing it right. The NV2X boards are the definitive example of doing the actual filtering properly, their only problem is their speed when doing so. The NV3X boards that we have seen to date and the R3X0 boards that we have seen do not do it right.
For that matter, if people really can't see the problems with ATi's AF implementation, why aren't they looking at nV's low quality high performing modes instead of their highest quality settings when comparing the R3X0 boards to NV3X offerings? Because different people see differing levels of issues on different levels. I can see the issues with the R300's AF implementation, although as I have stated previously(numerous times IIRC) their AA implementation is hands down better then nVidia's(again, why noone defending nVidia from my pokes at their issues?).
You show me a board with NV level drivers, NV2X quality AF, great levels of performance and R3X0 quality AA and I will buy it without hesitation(in that order in terms of priorities btw). That board does not exist right now, if S3 came up with it I'd buy it. If ATi came up with it I'd buy it. Hell, if 3dfx rose from the ashes and pulled it off I'd buy that too. My level of priorities has remained fairly constant for many years now, you can check through the archives and see I've been saying the same things(although my big push for hard T&L is now a dead point as everyone is doing it and very well right now) since these board were put in to place(longer then that actually, but these forums go back to '99 to frame where I'm coming from).
You swap my priorities around completely and you have the perfect candidate for a R3X0 based board.
I'd be curious to know what they are. I've not had a single crash since installing my 9700P. Didn't have any with the Ti4600 either. The only minor annoyance I've seen is the Half-Life thing when you use the escape key. I'm not playing it anymore, so it doesn't affect me. If I did,
~
quit
is all that is needed to get around it.
That's the issue(it happens in all HL powered games). Running under OpenGL it took my system down every time(Win2K SP3). Under D3D it just black screens, but running D3D you are limited to 12x9 res and have inferior performance. As far as using the console to escape, that works great if you are only trying to quit the game
That's the only thing that has managed to take my system down to date.
Ben, arent you making a big issue out of 1 issue with 1 game? You make it sound like there are such wide sweeping problems. Its a Half-Life engine game bug. You really should describe it as such since it is just that.
Isn't 1280 x 960 the res limit in OpenGL also? Is there a trick to get it higher? Yes, HL is played in OpenGL. I've never even tried D3D.
It's a hardlock bug, that isn't acceptable by today's standards. If it was a minor rendering error or some other such thing, that would be something else entirely. Also, it does impact one of the most popular games of all time, it isn't like it's a hard lock bug in an obscure title. If I was running a Win9X based OS I wouldn't even mention it, it was to be expected back then. Today, a bug that can hard lock a NT based OS is certainly a point that should be discussed, particularly considering all the hype around 'WHQL' drivers. I will gladly blast nVidia for the same thing if anyone could point me to a case where I can repeat the situation(hard lock) using a nV board.
Yes, there is a trick. Download the latest patch
They added 16x12 about a year ago IIRC I'm not saying its not a bug that doesn't need fixing. I'm saying it is one game (and mods). Your generalization of ATi drives locking up NT based systems seems purposefully blown out of proportion.Originally posted by: BenSkywalker
It's a hardlock bug, that isn't acceptable by today's standards. If it was a minor rendering error or some other such thing, that would be something else entirely. Also, it does impact one of the most popular games of all time, it isn't like it's a hard lock bug in an obscure title. If I was running a Win9X based OS I wouldn't even mention it, it was to be expected back then. Today, a bug that can hard lock a NT based OS is certainly a point that should be discussed, particularly considering all the hype around 'WHQL' drivers. I will gladly blast nVidia for the same thing if anyone could point me to a case where I can repeat the situation(hard lock) using a nV board.
Yes, there is a trick. Download the latest patch
Its one bug in one game
It's a hard lock bug, it's not like I'm talking about a minor glitch(I mentioned the fact that it was NT based as they don't suffer from mystery style W9X locks). I've listed games that have rendering errors, and Lithtech engined games that have performance issues, they are not in the same class as something that takes my system down. I'm used to measuring up time in months, and even then my reboots tended to be from driver updates(once in a while I would shut down). Currently I have two stand alone games installed(BS and HL- they do not rely on each other, and ignoring OpForce, CS and TFC) that I purchased that have issues with the bug.
If there was a bug that impacted the Quake3 engine and rendered all Quake3 powered games unplayable would it be worthy of making a big deal out of it? I have more HL powered games installed on my PC then I do Q3(by better then 2:1).
I also have HL/BS/OF. I'm not a CS player. I still view it as one game engine because...it is.
All I'm saying is you should refer to this problem as a lockup with HL engined games. Not generalize ATi drives as causing a system lockup. There are also people that don't run any HL engined games that are unaffected by this one bug. (I'm one of them).
If there were a Q3 bug...yes it should also be mentioned....as a bug when playing Q3 engined games. Same thing. Describe the issue accurately instead of generalizing it.
All I'm saying is you should refer to this problem as a lockup with HL engined games. Not generalize ATi drives as causing a system lockup. There are also people that don't run any HL engined games that are unaffected by this one bug. (I'm one of them).
If there were a Q3 bug...yes it should also be mentioned....as a bug when playing Q3 engined games. Same thing. Describe the issue accurately instead of generalizing it.
I wouldn't agree with that quote either. I found the nV drivers to be extremely good on both my GF3 and GF4. ATi has a ways to go before they can match them. nVidia has great driver support. I liked the 30.82s. It seemed the later drivers always had a lower level of IQ whenever I tried them. The poor AF performance of my GF4 was my only real gripe with it.
I still maintain the 9700/9800 are the best cards right now. The AA/AF performance AND QUALITY are better than anything else out there. The ridiculous heat/noise/cooling system of the 5800U made it a non option for me. From what I'm hearing so far, they have learned from this fiasco and the nV35 will not have this issue.
I'm looking forward to it.
I still maintain the 9700/9800 are the best cards right now. The AA/AF performance AND QUALITY are better than anything else out there. The ridiculous heat/noise/cooling system of the 5800U made it a non option for me. From what I'm hearing so far, they have learned from this fiasco and the nV35 will not have this issue.
I'm looking forward to it.
Ben
"The NV2X boards are the definitive example of doing the actual filtering properly."
I propose that an apples to apples comparision is almost impossible since both ati and nvidia use addaptive AF based on different principles. The problem? It's impossible to know exactly how much additional fill rate load the 9700 takes on when AF is enabled, since it's a function of the number of sloped surfaces in the scene, and the degree of each individual slope. And because nVidia uses its own adaptive algorithm, an apples-to-apples comparison between the two implementations where each both are doing equal filtering work is all but impossible.
SO
The most important way to evaluate the cards is to use them for yourself, this I have with my 2 NV20 cards a r200 and r300. I was not impressed with the NV20s AF and much prefered my 8500 and 9700. Now about the NV30 since I don't own it and have yet to see it run in reallife I suspend judgement.
Rogo
"The NV2X boards are the definitive example of doing the actual filtering properly."
I propose that an apples to apples comparision is almost impossible since both ati and nvidia use addaptive AF based on different principles. The problem? It's impossible to know exactly how much additional fill rate load the 9700 takes on when AF is enabled, since it's a function of the number of sloped surfaces in the scene, and the degree of each individual slope. And because nVidia uses its own adaptive algorithm, an apples-to-apples comparison between the two implementations where each both are doing equal filtering work is all but impossible.
SO
The most important way to evaluate the cards is to use them for yourself, this I have with my 2 NV20 cards a r200 and r300. I was not impressed with the NV20s AF and much prefered my 8500 and 9700. Now about the NV30 since I don't own it and have yet to see it run in reallife I suspend judgement.
Rogo
The NV2X is adaptive only in the manner in which AF is supposed to be adaptive, based on the angle of the textures. The NV3X and R3X0 are adaptive in terms of the amount of anistropy they apply no matter what the setting. If you have 8X AF set for the NV2X boards that is how they will sample all textures, the NV3X and R3X0 boards attempt to determine which textures need what level of anisotropy and it doesn't always work properly. The NV2X(and actually, the NV1X boards although they were limited to 2X) handle anisotropic filtering properly, the others don't.
Ben
I believe you're wrong about NV20s method.
It samples based on the slope of the textures, just like ati's BUT the footprint is adaptive meaning it can change from a rectangle to a trapezoid to a cross, depending on the x and y axis.
It doesn't filter the whole screen at the setting you've chosen. It's adaptive. I have a quote from David Kirk that explains this.
Rogo
I believe you're wrong about NV20s method.
It samples based on the slope of the textures, just like ati's BUT the footprint is adaptive meaning it can change from a rectangle to a trapezoid to a cross, depending on the x and y axis.
It doesn't filter the whole screen at the setting you've chosen. It's adaptive. I have a quote from David Kirk that explains this.
Rogo
Anything outside of the footprint is bilinear or trilinear filtered just like the r300 and nv30.
it doesn't filter AF on all slopes. That would actually be worse since the bandwidth required to filter unseeable textures would require a huge hunk of the availiable bandwidth, and I don't see why you'd think that would be "better" af.
Rogo
it doesn't filter AF on all slopes. That would actually be worse since the bandwidth required to filter unseeable textures would require a huge hunk of the availiable bandwidth, and I don't see why you'd think that would be "better" af.
Rogo
Post a link to the quote from D Kirk. nV was applying AF to lightmaps until very recently, and that has always been totally useless.
Originally posted by: Rogozhin
Claudii
"I play games obviously, but it's easier for me to tell which card has more accurate rendering from tests like this than screenshots of games on a website."
Rollo is a strange nut
But an intelligent one.
I don't automatically discount his pov as irrelevant but I do question his sense of "nvidia loyalty" that he seems to hold close to his heart.
Rogo
I can agree that synthetic testing methods are relevant to show differences in how each GPU does it´s rendering.
But I must also admit (also only having GF4MX420) that in games I (have to!) see no big problem. Of course I am not using any kind of AA, Quinqunx (who made that name??), SSx4/x8, or whatever (although it´s possible, not mentioning the nice slideshow). In some situations I use AF but with the GPU I have thats a long way from being much fun
Most games are action games where there isn´t much time to look at tearing lines or whatever artifacts one or the other GPU makes! I could of cause stop and just enjoy the landscapes, but I think I´ll die then...
I played Max Payne, NFSHP2, Unreal (the first one!, which I still think have OK graphics?), a cople of PinBallers, Serious Sam, FIFA 2003, TigerWoods 2002, etc. etc. and I can get the games to run quite OK for me for now, looking
OK on the screen...
BUT - and to me this is a BIG BUT(T!): nVidia with Mercedes Truck Racing, Midtown Madness, 3DMark2001SE - how come that nVidias make textures disappear?
This is STILL a big question mark for me??
PS: I like strange intelligent people, so nothing personal Rollo
Actually, the word predates the English language
It is an old Latin word that roughly tanslates to '5 into 1'.Your board OCd? I don't run in to that on 2K1SE(and my board is OCd), don't have the other two games insalled to test at the moment.
Yes, but even without OCing I don´t see the OilBarrel that pushes the truck aside in gamescene one. Not on my old TNT either.
It has really been a mystery, since it´s only in some resolutions (which i´m making a list of now).
As for the two games mentioned I think it is common known that nVidias don´t display correct...
It has really been a mystery, since it´s only in some resolutions (which i´m making a list of now).
As for the two games mentioned I think it is common known that nVidias don´t display correct...
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter