Genx87-
While the NV30 had obvious issues with balanced and aggressive the application mode is much better than the highest quality on the 9800 Pro. Same can be said of the 5900 Ultra.
From what I've seen of the newest drivers I would certainly say this is in question as of now. nVidia's AF on the NV2x was a staggering amount better then that on the R2x0 boards. For nV, the problem is that they declined significantly while ATi improved significantly. The R3x0 still can't touch the AF quality that the NV2X boards have, but neither can the FX. None of the latest gen parts have what I would call very good AF, nVidia's was at least better then ATi's but the latest drivers are showing that that may no longer be the case. This is one of the reasons I'm hoping that Volari will turn out to be better on the driver and filtering front then either of the current boards, they simply don't cut it.
BFG
Then why hide something that they're not doing?
I've already explained this. nVidia has a history of hiding what their drivers are doing since the inception of the company. That is why there has never been a fully functioning open source nV driver. They like to keep what they are doing hidden, look at the amount of patents that they have.
think they're doing it to inflate performance which is why nVidia chose to follow their example. ATi and 3dfx obviously quite rightly felt image quality is more important than providing an unusable feature on their cards.
On the high end you would see a ~10% performance difference using S3TC1 versus S3TC3. That is with a lower level of compression. Why do you think they would do something that may increase the performance by maybe 5% in a bench they already dominated?
Because it's turned off by default.
Not in JKII and not in JKIII, have to double check on RtCW and Alice to be sure on those. The JediKnight games it is certainly enabled by default, certain level textures are too much for 128MB boards without TC on(the one you have to sneak around in on JKII as an example).
Also do you have Medal of Honour?
Was going to buy it, played it for about forty five minutes and had to wash my mouth out afterwards, couldn't believe how insanely linear it was(no, never purchased it in case you didn't figure ).
That's still cheating. If the user requests trilinear AF in the drivers then the drivers have no right to override those settings.
Then ATi is cheating and has been for years. They do not proper filtering implementations either(AF), and they haven't since the R200 debuted. There AF has always been adaptive and with the R200 it also forced bilinear no matter what. If you say nV is cheating on this one, then you have to say ATi is cheating and has been for years. I haven't said ATi was cheating on their filtering, just that it sucked. From what I've seen of the FX with the latest drivers, I'd have to say it looks like it is sucking pretty bad too(although it still doesn't look nearly as bad as the R2X0, it would be hard to be worse then that- and I still didn't call that cheating).
Application detection is bad because it's a fragile optimisation that can break with later versions of the program or by simply renaming the executable for whatever reason. It also removes all control from the user.
App detection nearly never breaks the game with later revisions, if it did you would have had Kyro owners frothing mad at the amount of games that wouldn't work. As far as taking control away from the user, you have never really had full control over your video card. I'm with you that I want as much control as possible, but you aren't getting it with your 9700Pro, you didn't get it with your GF4 prior to that or any of the boards you have owned since you have been here.
That gives control back to the user instead of just having the drivers to run off and do whatever they like without telling anyone. It also takes away the blame of cheating from the manufacturer since the reviewer can untick it if he/she doesn't like what the option does.
It is nothing new, and overwhelmingly you don't want users to have that kind of access to control. You or I or the overwhelming majority of AT users would be perfectly fine with that level of control, most people wouldn't. Can you imagine the support nightmare it would create having 'I broke my cupholder' users being allowed to assign the Zbuffering properties or caching limitations on their video cards? Of course applying simple filtering implementations is a basic request, but its not like you are getting pure filtering from ATi now either. For that matter, how does your R9700Pro overclock using the default BIOS and drivers? All of the vendors have certain things that they should implement IMO but don't. I am fully with you on the FX going to far right now, that's why I won't buy one.
And they did, with a mixed mode, hand-optimised, Microsoft compiled rendering path. If that path can't beat ATi's then there's no way the full precision path is going to do so, with or without Microsoft's compiler.
Actually with the mixed mode and the latest drivers it appears that nV does edge out ATi in some of the benches. But that is neither here nor there. The real reason they did it is for people like you. You now are under the impression that ATi's boards absolutely destroy the FX boards with PS 2.0 performance. That is not the case. ATi has an edge, but if the actual ~25%-35% was shown instead of the 70% they were able to rig up where would all the ATi hype be?
Why waste time and resources on something as useless as that?
They compiled it for ATi to the compiler that best matched their hardware, but they failed to do it for nVidia. They wanted to show a larger performance difference then there actually is(as Dave mentioned, it was to show the shader performance difference), and they have some people buying it. Look at Halo performance with PS 2.0, real staggering difference there isn't it? We have one top notch title shipping with PS 2.0 support and it doesn't show anything remotely like what we saw with HL2, don't you find that the least bit odd?
At some point you just have to accept bad hardware when you see it and move on.
Gearbox, who wasn't paid millions in promotional money from ATi, managed to get decent PS 2.0 performance out of this 'bad hardware'. Doesn't that ring any bells for you? If one company that was paid millions by company X comes out and shows how fast company X is, while another company who wasn't paid millions of dollars shows a finished shipping product that isn't remotely close to showing the same thing it doesn't imply to you that perhaps you shouldn't jump to conclusions quite yet?