New dynamic branching demo

Not entirely wrong. HardOCP, this site, elite bastards, etc were all saying that SM3.0 would only offer performance increases, and that everything in SM3.0 could be done in SM2.0. This was regurgitated over and over. Now it appears a SM3.0 game is coming out that has significant IQ increase as well - regardless of the reason, people without SM2.0 may be losing out on visual effects. That is unexpected and could annoy quite a few X800 buyers.

The branching effect also exhibits the 10% slowdown on 6800 series cards. Nvidia cards get a performance *hit* when Humus' "branching" is enabled, not a gain. Has nothing to do with the card, but the coding.

 

right becasue it can't be a driver problem ?

 

If it was there was obviously no attempt to work around it or test it, or get it working etc... His last demo also had problems on Nvidia cards. This furthers the idea of being ATI biased

 

are you kidding me?

lets compare the number of sm3.0 cards to 2.0 cards shall we...

sm3.0 - nv4x
sm2.0 - any dx9 card which includes the nv4x

now lets assume that "SM3.0 effects would most likely run too slow on any SM2.0 card with the exception of the X800 series based on the performance of the beta patch at E3" is true...

guess what, the nv4x can run sm2.0 as well, so if they did incorporate a technique similar to humus', you would be coding for 2 cards

...so yes, "how worthwile it is to code for one card series is questionable"

 

And why are you so "angry"?
If a person with a X800 don't get this SM3.0 in games it's fine...because they bought a card with no SM 3.0
But, if you bought a 6800, and a guy with a X800 can run the same SM3.0 stuff that you can, the fool here will be you, and nVidia PR machine,not the guy with a X800.
We must wait for a real SM3.0 game to claim the winner in this round...last time i believe in nVidia PR i got a crapy FX5900 while my friend got a 9800. I can't run Farcry in all it's glory, but my friend can, with a card that don't have all these PS2+++++ nVidia claims to have.
Let's wait, and let's hope that this Humus trick work.

 

well humus does include the source code to all his demoes, if you find something wrong with his coding, you can fix it yourself or discuss with him about it

not a programmer? well then, it sort of defeats your argument

 

Humus, do you have an nv4x? if so, could you code the same demo using SM3.0 for comparison? or am I just talking nonsense and the workaround code you wrote would appear the same for SM3.0...

 

Yes, and it's the truth. This has nothing to do with the technique only being usable in PS3.0. As shown, it can be done in PS2.0. This is a marketing deal, and it would be no different from, say, Half-Life 2 auto-detecting Nvidia cards and automatically degrading texture quality to lowest settings.

 

But it doesn't have any IQ increases that couldn't be done just as easily with SM2.0, I think that was sort of the reason for Humus' demo...to show that there was an easy alternative and to show that the ONLY reason FarCry will be SM3.0 only is $$$$.

 

Not regardless of the reason, it should bother both 98XX & X800 owners because there will be effects implemented which their hardware is capable of and yet aren't available to them for one simple reason - TWIMTBP. Crytek's game is meant to be played on all PCs, not just the one's carrying the current gen. of Nvidia's cards. It is their responsibility to implement an SM 2.0 fallback, otherwise they are alienating 99%+ of their customers unfairly. I don't see how any current FX59xx owner can know this and not be pissed. They're being told that even though they spent $300 - $500 within the last 12 months they will not be able to experience effects their hardware is capable of, albeit @ lower rezes, just because Nvidia & Crytek feel like it.

 

humus may not have access to the cards . He is doing it on his own time. If you find problems you should send them to him or post them here .

Its hard to code to make sure something works with so many diffrent driver sets when u don't have access to the cards .

Hardocp , this site , elite baters and all were correct.Sm 3.0 would only offer performance increases .

It now apears that nvidia has payed a company not to code the new effects on sm 2.0 and instead make them a sm 3.0 only feature .

Yes x800 , r3x0 , nv3x people are going to be very annoyed . But they should be annoyed at nvidia and crytech not at ati .


Like i said . If half life 2 would only support p.s 1.1 on 6800ultra cards nvidia owners would be very pissed.

You could not blame nvidia as they supported p.s 2.0 / 3.0 . You would have to be pissed at ati and valve .

There is no diffrence.

Its just that right now your fav company has the support. If it was reversed i'm sure you would be pissed.

 

I think this is getting out of hand.

question is not whether SM3.0 can be emulated by SM2.0. yes
Humus showed it can be done certain extent.

but almost anything can be emulated by turing machine
or in GPU case minimum graphical instruction set.

important issues is whether SM3.0 will reduce complexity and easier
for developers to code. hence it is possible to code large complex
shaders and possibly increase functionality / performance with SM3.0
rather than 2.0.

so in order to true comparison, what is complexity of coding this
in SM3.0 vs SM2.0 in non trivial game? how easy to maintain,
how easy to tune?

 

Personally, I wouldn't judge a game by it's optimizations. I don't want companies to make a decision for me. I think Half-Life 2 will be a great game even if it only allows 6800U for SM 1.1(in which case I will buy an X800 Pro for sure).

I wouldn't call this technique dynamic branching. I would call it a replacement for one of dynamic branching's advantages.

 

holy crap...

9500 pro - 1024 x768

false - 20-30 fps avg
true - 70-100 fps avg


million dollar question... could this be applied to current game's lighting like far cry, or forth coming doom 3 for a speed up?

 

Far Cry 1.2 contains no PS3.0-specific effects, only performance benefits resulting from the use of PS3.0. And even with that, you need a DX9.0c beta installed.

 

Ummm, while I can't remember every one of our sites comments, I don't believe we've ever explicitly stated such a thing.

 

The reasons are important. You are strongly implying that there is some technical reason why SM2.0 could not do these effects. In fact there is no technical reason why the same visuals could not be supported under SM2.0. The higher clock rate of R420 would go a long way towards offsetting any effeciencies from using SM3.0 over SM2.0.

If it's true, the only reason Crytek has not done these effects in SM2.0 is because Nvidia bribed them not to, thus orphaning all SM2.0 cards, including Nvidia NV3x customers. Just as Nvidia has been attempting to bribe other developers to do the same.

This isn't about the capabilities of the cards, it's about Nvidia's ability and willingness to spend money to screw it's customers and that of it's competitors by bribing developers. It's this same kind of strongarm/bribary technique that put the likes of Microsoft on the wrong side of the US anti-trust regulators in recent years.

 

My God, it's full of stars.

 

Well I'm sure someone did in your forums so it's the same difference.

Baron- "Open the pod bay doors Hal."

 

Maybe because "HDR" is not a single, well defined effect, but actually a whole range of things (no pun intended). Radeon cards can do some of it, but not everything. NV40 can do some more.