GTX295 vs 4870X2

[Keysplayr]

"Right, but how many nVidia hardware accelerated PhysX games are there? UT3?s five maps that nobody plays don?t count, and neither do tech demos, nor do games that will never be released to the public."

You only have about two to three weeks left for this argument. So use it to the hilt while you can.

BTW, you can talk polygons and stencils shadows and first class polygon edges and flux capacitors for that matter until doomsday.
Under a magnifying glass, BFG, you probably could tell a difference from 4x to 8x AA. But can you actually tell the difference while gaming? Especially while gaming? Of course you can't. Can you tell the difference between 0x and 4x? Absolutely. You'd notice crawl everywhere at 0x. And you of all people should understand where I'm coming from when I say that it is even less noticable at higher resolutions. You, who considers 1920x1200 a "middling" resolution as we have discussed many times before. Especially in a moving game.

Screenshots are good for close minutia inspection for super subtle differences when they are magnified. In fact, most reviews that compare AA, that I can remember, all use still screenshots under magnification for comparison.

I wish I had the monetary resources and time. Id fly you in. I'd set up a delightful test room for you with 10 computers all lined up in a row. All identical on the outside down to the mousepad. They might even have all the same hardware inside, or they might be totally different. You would never know the hardware, or the settings used in each. Same game and level on each system. I'd invite you to play that game on all 10 systems with different levels of AA on each. Nothing below 4x, and nothing above 8x. I could very easily say, that you would not think it so easy to discern which systems utilized 4x or 8x AA with your naked eye and whilst playing. There is no way. I'd only publish your score at the end of the session with your permission. . But you might not be so eager to give it.

Do you think you could accomplish something like that? Guess which level of AA each system is using after playing on them? Go ahead, tell me you could.

 

[dadach]

good one BFG...its rather funny to see nvidia payroll guys squirming and downplaying the facts you presented them with, which they have no answer for

no matter what...all the reviewers need to explore every single available setting in drivers...im sure once the REAL reviews start coming out, we will be smarter about how much this new card is better than atis old card

 

[Keysplayr]

No dadach. It was someone else who had the, "good one".

 

[BFG10K]

Originally posted by: keysplayr2003

You only have about two to three weeks left for this argument. So use it to the hilt while you can.

What happens in 2-3 weeks? A flood of PhysX titles that will saturate the market? :roll:

Under a magnifying glass, BFG, you probably could tell a difference from 4x to 8x AA. But can you actually tell the difference while gaming? Especially while gaming? Of course you can't.

Um, during gaming is exactly where the difference is more visible ? when the scene is moving.

Still screenshots are used because they?re more accessible to readers than HD video, and also because it?s much easier to zoom the image rather than trying to get video playback to cover exactly the same areas. It?s a logistics issue rather than anything else and that?s exactly why screenshots require zooming, because they?re unsuitable for the task they?re being used for.

I mean let?s step back a bit and think about your comments. If anything over 4xAA was useless, why would nVidia invest millions of dollars in R&D to equip their hardware to handle 8xMSAA along with the new CSAA modes? What an utter waste of time and money using your reasoning.

Way back in the Radeon 9700 Pro days it was proven that 6xMSAA provided a visible benefit over 4xMSAA, so is it really that surprising that 8xMSAA also provides a benefit given it?s even stronger?

Do you think you could accomplish something like that? Guess which level of AA each system is using after playing on them? Go ahead, tell me you could.

Absolutely not, but depending on how good your test was I should be able to detect the systems that are running more than 4xAA.

 

[Keysplayr]

Originally posted by: BFG10K

Originally posted by: keysplayr2003

You only have about two to three weeks left for this argument. So use it to the hilt while you can.

What happens in 2-3 weeks? A flood of PhysX titles that will saturate the market? :roll:
What happens in 2-3 weeks is, you can't say there isn't any PhysX titles on the market. Because there will be the first, we'll second actually if you count warmonger, but the first big game to enter the retail world with PhysX. A flood is not necessary to render your rant moot.

Under a magnifying glass, BFG, you probably could tell a difference from 4x to 8x AA. But can you actually tell the difference while gaming? Especially while gaming? Of course you can't.

Um, during gaming is exactly where the difference is more visible ? when the scene is moving.
Yes, when going from 0AA to AA.

Still screenshots are used because they?re more accessible to readers than HD video, and also because it?s much easier to zoom the image rather than trying to get video playback to cover exactly the same areas. It?s a logistics issue rather than anything else and that?s exactly why screenshots require zooming, because they?re unsuitable for the task they?re being used for.

I mean let?s step back a bit and think about your comments. If anything over 4xAA was useless, why would nVidia invest millions of dollars in R&D to equip their hardware to handle 8xMSAA along with the new CSAA modes? What an utter waste of time and money using your reasoning.
But you don't understand. This is not about ATI/Nvidia. This is about noticing a difference from 4xAA to 8xAA while playing. Money spent to further technology on either side is irrelevent. Both companies continually develop their technologies, isn't that so? Would you have them stagnate?

Way back in the Radeon 9700 Pro days it was proven that 6xMSAA provided a visible benefit over 4xMSAA, so is it really that surprising that 8xMSAA also provides a benefit given it?s even stronger?
This is not way back in the 9700Pro days.

Do you think you could accomplish something like that? Guess which level of AA each system is using after playing on them? Go ahead, tell me you could.

Absolutely not, but depending on how good your test was I should be able to detect the systems that are running more than 4xAA.

LOL, how good my test is? All other things being equal, it's either 4xAA or 8xAA. Done. And no, there is no way on this earth you can tell me the correct level of AA being used on all ten of those computers. Unless you are the luckiest person on earth, in which case you should play the lottery. Because that is all you could do. Give your best guess, nothing more. So, it's not that you "should" be able to detect the systems running more than 4xAA, it's how many times you "guess" correctly. End of story.

 

[nRollo]

Originally posted by: BFG10K

Why screenshots? Anyone that actually uses AA knows the real benefit comes in actual gameplay during moving scenes.

Are you saying that the reviewers don't use AA, or that I don't? I'd have to disagree with that. I've posted links that support my position, can you post links that support yours please? I'm open minded, and if you post links to videos that illustrate this large difference you say exists, I'll retract. Heck you can even post them yourself on ABT if you want, then we can all watch and decide for ourselves.

Originally posted by: BFG10K
I mean if we?re going off screenshots, can we compare ATi?s AF to nVidia?s AF with screenshots and conclude the cards are equal? Of course not. In fact (and rather ironically) ATi?s AF often looks ?better? in screenshots but in actual gameplay that isn?t the case at all.

Personally I'd say ATi and NVIDIA AF are pretty close to equal- are you saying that you believe NVIDIA's is superior? I realize NVIDIA's is closer to angle invariant, but you're the first person I've seen say there's a noticeable difference. Seems like something I would have likely been posting if this were widespread knowledge- source?

Originally posted by: BFG10K
There are visible and clear improvements in actual gameplay when going from 4xMSAA to 8xMSAA. That is a fact.

You and others have been saying that, can you post links to your source, or better yet videos that illustrate? I only remember reading the difference from 0X-4X is large, and 4X-8X is small. As it's impossible for CSAA to have worse AA than 4X (sample patterns: 4XMSAA = 4/4/1, BX MSAA =8/8/1, and 8X CSAA = 8/4/1) it seems likely my theory that 8X CSAA is somewhere between 4X and 8X in quality. People should note that although 8X MSAA offers double the coverage and color sampling, I've never seen anyone put forth it offers double the image quality. The jump from 4X to 8X to 16X is seen as diminishing returns- do you have contrary info?

Originally posted by: BFG10K
Right, but how many nVidia hardware accelerated PhysX games are there? UT3?s five maps that nobody plays don?t count, and neither do tech demos, nor do games that will never be released to the public.

How do you know nobody plays them? I'd bet $100 more people play them than the 8 year old games I had never heard of you were upset about on nZone, or Serious Sam 1 with a specific patch you were after NVIDIA to fix. Unless you know everyone or have some method of monitoring them, you can't say this.
What abnout GRAW2 and Warmonger? They are games people can play now. Mirrors Edge launches next month, Cryostasis in February, do you have information they won't? Or information about all of the other upcoming games? (e.g. Bionic Commando)

Originally posted by: BFG10K
Now compare your list to the thousands of 3D games that can have 8xMSAA enabled.

How about I compare it to DX10.1 or tesselator enabled games instead, as those are vendor specific features the X2 has?

Better yet, how about link us to games where 8X MSAA is less playable on a GTX200 SLi rig than a 4870X2, as you seem to be implying the 8X MSAA is a reason to choose ATi over NVIDIA.

I say the IQ benefit is small and offset by better transparency and CSAA, and that this card will offer 8X MSAA at a level the difference in ATi and NVIDIA won't matter.

Please provide links that disprove what I say, rather than just saying I'm wrong and pointing at your single card results which aren't applicable. You could get both of the cards and review them for ABT perhaps.

Originally posted by: BFG10K
I have about 80-85 games installed right now where I can enable 8xMSAA. I have two PhysX titles and neither of them support hardware acceleration, so hardware PhysX makes no difference to them. So remind me again, why should I be excited about hardware PhysX?

There's no reason for you to be, the people who own the current games and will be buying the games launching next year are the ones who should be excited. I don't own any sailboats, only have powerboats and hand powered boats, but I don't say "Why should I be excited about a nice sailboat?" and totally discount their value.

Originally posted by: BFG10K
Show me the majority of new games (not tech demos) shipping with hardware PhysX support and then I?ll get excited enough to install the PhysX driver. Until then it?s a niche at best, like DX10.1.

OK so niche products aren't worth having to you. Ironically you made a point about no TSAA in OpenGL, and that's a niche these days. So I guess to you OpenGL is the niche to care about, PhsyX is not. Whether you care about PhysX should be irrelevant to everyone else presumably- it offers higher image quality and changes gameplay in games current and upcoming. The fact they aren't the majority of games doesn't change anything. Used to be 3d games were few and far between, made most people want to play them all the more.
Were you saying "There's only a handful of 3d games, I'll stick with good ol' bitmaps"?

Eh? Since when are Call of Duty 4 or Bioshock five year old games?

When you show me a GTX295 can't run these games acceptably at the resolutions a X2 can, and answer my previous questions, we'll talk. Maybe I'll post some benches as I'm in a good position to do so. (hint)

Originally posted by: BFG10K
So let me see if I read this correctly: you?re saying 8xMSAA doesn?t make much difference but 8xCSAA does, even though it?s objectively and mathematically provable that 8xCSAA is inferior to 8xMSAA?

I didn't say that at all. I said 8XAA offers a small improvement over 4X, that 8X CSAA offers a smaller one, but the difference in either and 4X isn't enough to get worked up about.

Originally posted by: BFG10K
That simply doesn?t fly. If 8xMSAA provides no benefit then neither does 8xCSAA, making it a useless feature. But if 8xCSAA provides a benefit then 8xMSAA must also provide one because it?s objectively provable that it?s better.

You've said this many times in this post, I posted links to others who agree with me and screenshots to back my position. I ask again, can you do the same? I also never said it's "useless", please don't misquote me. I said the difference in 4X MSAA and 8X MSAA is much smaller than the difference in 4X and 0X, that 8X CSAA provides some of the small improvement by having double the coverage samples.

Originally posted by: BFG10K
I have no trouble acknowledging the strengths of either vendor which is why I don?t make ridiculous flip-flop arguments depending on which IHV is currently leading or what is the current ?in? thing. I?ve extolled the virtues of nVidia?s combined AA modes for years despite the fact that basically no-one uses them.

OK sounds good.

Originally posted by: BFG10K
OpenGL games don?t need to launch; there?s already an installed base going back over ten years to GLQuake. Additionally there are large community followings of source ports such as Doom, Quake, Quake 2 and Descent 1/2. Additionally, many popular legacy games like UT99 have updated renderers that run best on OpenGL.

PhysX needs launches because it needs games for it?s almost non-existant market base, namely titles supporting hardware acceleration on nVidia GPUs.

Everything starts somewhere BFG, and there are games now and many on the way in 2009.

Originally posted by: BFG10K

Again with the still screenshots which are generally useless for showing IQ differences unless they?re zoomed. Anyone that uses AA in gaming knows that.

Ratchet doesn't know what he's doing? Have you told him this over at B3d? In any case, I'm just waiting for you to post links to contrary proof still, because I've used 8X AA and have missed the awe inspiring difference you seem to be alleging and the reviewers always miss.

Originally posted by: BFG10K
8xCSAA?s effectiveness diminishes on polygon intersections and it?ll also fall back to the base 4xMSAA on stencil shadows, when running TRAA, and also anywhere where coverage sampling may fail for a given pixel.

8xCSAA is definitely better than 4xMSAA but 8xMSAA is even better. The only place where 8xCSAA may rival 8xMSAA is on first-class polygon edges and only if all pixels succeed in their coverage tests.

I understand there are differences and fallback with CSAA, and realize MSAA is by definition "better", but think it's a diminishing return compared to 0X-4X, and not a whole lot better than 4X.

I also say NVIDIA GT200 SLi solutions offer good 8X AA performance, sometimes better, and the small difference in IQ and small performance lead in some games doesn't justify anything. I also say it's offset by transparency deficiencies for ATi.

 

[BFG10K]

Originally posted by: keysplayr2003

What happens in 2-3 weeks is, you can't say there isn't any PhysX titles on the market. Because there will be the first, we'll second actually if you count warmonger, but the first big game to enter the retail world with PhysX. A flood is not necessary to render your rant moot.

So two games suddenly make PhysX relevant? LMAO.

So therefore using your logic, DX10.1 is relevant because two titles - Far Cry 2 and Stalker Clear Sky - are already in retail and support it?

Face it, PhysX is about as relevant as DX 10.1. If you consider one an advantage because of upcoming titles then you must also classify the other in the same manner.

Yes, when going from 0AA to AA.

Yes, but that?s not the only case.

But you don't understand. This is not about ATI/Nvidia. This is about noticing a difference from 4xAA to 8xAA while playing. Money spent to further technology on either side is irrelevent. Both companies continually develop their technologies, isn't that so? Would you have them stagnate?

If there's no difference going above 4xAA you must be claiming both vendors are wasting time and money developing such AA methods. Is that what you're saying? It?s a very simple question to answer, so please answer it.

Why develop such AA methods if, according to you, they?re useless?

This is not way back in the 9700Pro days.

What an absolutely ridiculous answer. Do you feel the principles of MSAA have changed since 2002?

LOL, how good my test is? All other things being equal, it's either 4xAA or 8xAA. Done. And no, there is no way on this earth you can tell me the correct level of AA being used on all ten of those computers.

Like I said, if your test is setup correctly allowing an identical side-by-side comparison of actual in-game movement then I wouldn?t have too much trouble doing it. It?s really not that hard once you?ve trained yourself to spot aliasing artifacts like I have.

I mean using your scenario most wouldn?t be able to tell the difference between the vendors? anisotropic filtering, but you don?t see me running around and claiming they?re equal. I can easily spot the difference in actual gameplay.

 

[Keysplayr]

"It?s really not that hard once you?ve trained yourself to spot aliasing artifacts like I have."

Bingo!!! So one has to train oneself to actually have a chance of noticing differences.

This is perfect, thank you.

 

[BFG10K]

Originally posted by: nRollo

Are you saying that the reviewers don't use AA, or that I don't? I'd have to disagree with that. I've posted links that support my position, can you post links that support yours please?

No, I?m saying screenshots are generally useless for showing AA differences and hence they require zooming. I?m also saying in-game movement is the ideal test-bed because aliasing is more apparent in motion.

As for your link: http://episteme.arstechnica.co...002061831#456002061831

It?s painfully obvious how inadequate 4xMSAA is compared to the higher modes, and again those are still screenshots which aren?t even zoomed.

Heck you can even post them yourself on ABT if you want, then we can all watch and decide for ourselves.

Sorry, self-promotion is against the forum rules so I can?t do that. The link to Ars above will have to suffice.

Personally I'd say ATi and NVIDIA AF are pretty close to equal- are you saying that you believe NVIDIA's is superior? I realize NVIDIA's is closer to angle invariant, but you're the first person I've seen say there's a noticeable difference. Seems like something I would have likely been posting if this were widespread knowledge- source?

You aren?t posting accurate information about AA so why should you post accurate information about AF?

They?re visibly different - during in-game movement you?ll find ATi shimmers and crawls much more than nVidia does. As for links, sorry, I?m not allowed to self-promote.

People should note that although 8X MSAA offers double the coverage and color sampling, I've never seen anyone put forth it offers double the image quality. The jump from 4X to 8X to 16X is seen as diminishing returns- do you have contrary info?

I never claimed it offers double the image quality because that?s really not possible unless you have a woeful pixel pitch. There?s certainly the law of diminishing returns after 4xAA but to say there?s no difference at all is false. In fact that jump from 4xAA to 8xS can be massive, thought technically we?re talking about more than MSAA in that case.

How do you know nobody plays them? I'd bet $100 more people play them than the 8 year old games I had never heard of you were upset about on nZone, or Serious Sam 1 with a specific patch you were after NVIDIA to fix. Unless you know everyone or have some method of monitoring them, you can't say this.

You must be joking. Who do you think plays those 5 UT3 maps? Seriously, who?

What abnout GRAW2 and Warmonger? They are games people can play now.

Right, just like the dozens of OpenGL games people can play now, and do.

How about I compare it to DX10.1 or tesselator enabled games instead, as those are vendor specific features the X2 has?

Why? This whole topic started because of you downplaying 8xMSAA. Don?t change the subject.

Better yet, how about link us to games where 8X MSAA is less playable on a GTX200 SLi rig than a 4870X2, as you seem to be implying the 8X MSAA is a reason to choose ATi over NVIDIA.

I?ve already posted links in the past where 8xMSAA is slower on my 260+ than it is on the 4850.

There's no reason for you to be, the people who own the current games and will be buying the games launching next year are the ones who should be excited. I don't own any sailboats, only have powerboats and hand powered boats, but I don't say "Why should I be excited about a nice sailboat?" and totally discount their value.

Exactly the same applies to DX 10.1.

Anyway, I own current games and I will be buying games that launch next year. That still doesn?t make me excited about PhysX.

OK so niche products aren't worth having to you. Ironically you made a point about no TSAA in OpenGL, and that's a niche these days. So I guess to you OpenGL is the niche to care about, PhsyX is not. Whether you care about PhysX should be irrelevant to everyone else presumably- it offers higher image quality and changes gameplay in games current and upcoming.

No, you?ve come full circle again. The original point was you discounting AAA in OpenGL but then I pointed out there were more OpenGL titles than PhysX ones.

When you show me a GTX295 can't run these games acceptably at the resolutions a X2 can, and answer my previous questions, we'll talk. Maybe I'll post some benches as I'm in a good position to do so. (hint)

So you?re playing the playable game again, where benchmark victories don?t matter as long as both vendors are playable?

Ah yes, the same old Rollo. Just like in the 5800 Ultra days where it didn?t matter the 9700 Pro was faster in UT2003 as long as both cards were playable. :roll:

I didn't say that at all. I said 8XAA offers a small improvement over 4X, that 8X CSAA offers a smaller one, but the difference in either and 4X isn't enough to get worked up about.

I disagree with that, and presumably nVidia does too given they invested R&D to add the capability to the hardware.

I also never said it's "useless", please don't misquote me. I said the difference in 4X MSAA and 8X MSAA is much smaller than the difference in 4X and 0X, that 8X CSAA provides some of the small improvement by having double the coverage samples.

You?ve repeatedly implied 8xMSAA is a non-issue because nVidia has 8xCSAA. This is just plain false.

Ratchet doesn't know what he's doing?

Ratchet posting a screenshot doesn?t magically remove the inherent limitations of a screenshot.

But since you?re so insistent: http://techreport.com/articles.x/11211/8

You can see 8xQ providing a visible gain in IQ over 4xAA and 8xAA.

I also say it's offset by transparency deficiencies for ATi.

Except nVidia?s TrAA doesn?t work in OpenGL (sadly).

 

[emilyek]

Stupid question:

Can you SLI these?

Could you do 3-way SLI with these on a ASUS P6 Revolution board since it has 6 lanes?

 

[BFG10K]

Originally posted by: keysplayr2003
"It?s really not that hard once you?ve trained yourself to spot aliasing artifacts like I have."

Bingo!!! So one has to train oneself to actually have a chance of noticing differences.

This is perfect, thank you.

If you somehow think you?ve ?got? me remember, there are people out that that don?t even notice if AF and AA are disabled.

The fact is you?ve trained yourself to detect the difference between 0xAA and 4xAA, whether you admit it or not.

Now imagine those same people asking you to check ten rigs for AA levels like you asked me.

 

[Keysplayr]

Originally posted by: emilyek
Stupid question:

Can you SLI these?

Could you do 3-way SLI with these on a ASUS P6 Revolution board since it has 6 lanes?

I think the maximum is two. Quadfire or Quad SLI.

 

[chizow]

Originally posted by: BFG10K
So two games suddenly make PhysX relevant? LMAO.

So therefore using your logic, DX10.1 is relevant because two titles - Far Cry 2 and Stalker Clear Sky - are already in retail and support it?

Face it, PhysX is about as relevant as DX 10.1. If you consider one an advantage because of upcoming titles then you must also classify the other in the same manner.

No, they're not even close really. Even the staunchest ATI supporters have dropped this line of debate as its painfully clear that DX10.1 is insignificant, especially given Nvidia DX10 parts support the only implemented DX10.1 feature anyways.

Its been over a year since DX10.1 launched and all we have to show for it is 2.1 titles that support it. Actually 1.1 since it seems Far Cry 2 running in DX10.1 just results in horrible stuttering that ATI has no fix for. The saddest part is DX10.1 doesn't even provide the eye-candy enhancements PhysX is derided for.

PhysX on the other hand has shown significant gains in the 10-month period since Nvidia acquired Ageia. CUDA PhysX and legacy title support 4-months later. Updated Power Pack 3 months later. Mirror's Edge and Cryostasis gameplay demos recently. We'll see a glimpse of what PhysX can do in games developed with it from the ground-up in the next two months with Mirror's Edge and Cryostasis.

Here's another title showing just how much difference PhysX can make. I don't think anyone can argue with any credibility that DX10.1 or even 8x over 4x MSAA can result in such stunning visual differences.
Cryostasis Gameplay Trailer

 

[Creig]

Originally posted by: chizow
PhysX on the other hand has shown significant gains in the 10-month period since Nvidia acquired Ageia. CUDA PhysX and legacy title support 4-months later. Updated Power Pack 3 months later. Mirror's Edge and Cryostasis gameplay demos recently.

You call the current PhysX games list significant? Nvidia may be working on PhysX, but so far they have precious little to show for it.

Originally posted by: chizow
We'll see a glimpse of what PhysX can do in games developed with it from the ground-up in the next two months with Mirror's Edge and Cryostasis.

The Nvidia crowd keeps repeating "just wait, you'll see" in regards to PhysX support. Well, you're right. We're waiting.

 

[chizow]

Originally posted by: Creig
The Nvidia crowd keeps repeating "just wait, you'll see" in regards to PhysX support. Well, you're right. We're waiting.

Haha well there's the staunchest ATI supporters, and then there's Creig.

My point was that if PhysX is insignificant, then DX10.1 is even less significant. You simply can't compare the two as has been shown in the few glimpses we've seen. Even in its most basic form, eye-candy special effects, it completely blows away anything DX10.1 offers over DX10.

First post-CUDA PhysX title launches Jan 6th with Mirrors Edge. Are you and others with ATI parts really waiting? I don't think so.

 

[Keysplayr]

Originally posted by: Creig

Originally posted by: chizow
PhysX on the other hand has shown significant gains in the 10-month period since Nvidia acquired Ageia. CUDA PhysX and legacy title support 4-months later. Updated Power Pack 3 months later. Mirror's Edge and Cryostasis gameplay demos recently.

You call the current PhysX games list significant? Nvidia may be working on PhysX, but so far they have precious little to show for it.

Originally posted by: chizow
We'll see a glimpse of what PhysX can do in games developed with it from the ground-up in the next two months with Mirror's Edge and Cryostasis.

The Nvidia crowd keeps repeating "just wait, you'll see" in regards to PhysX support. Well, you're right. We're waiting.

Why are you waiting for PhysX?

 

[chizow]

Originally posted by: keysplayr2003
Why are you waiting for PhysX?

He's trying to make it sound like pie-in-the-sky type stuff like DX10.1

 

[WT]

I waited a good 5 months to buy my video card for my planned summer upgrade, as it seemed better to wait until ATI had released the 1gb 4870, but then after its release the GTX260 216 card kept me waiting yet again, until I finally caved in November and ended up buying the GTX280, of all things.
Reading up on that, it would appear that my Q6700@3.5 ghz is still a bottleneck for the system @ 19200x1200, so I guess my BFG Step Up is not a worthwhile option for this particular card. It looks nice on paper, but spending $100 for no noticeable gain is just not money well spent in my case.

But .. damn it .. I've bought plenty of hardware with no apparent gain many a time previously, so don't count me out yet !!

 

[Creig]

Originally posted by: chizow
Haha well there's the staunchest ATI supporters, and then there's Creig.

Maybe you should leave the personal comments out of your posts and stick to the thread topic. FYI, my last card was an XFX G92 8800GT.

Originally posted by: keysplayr2003
Why are you waiting for PhysX?

To see if it:

A) Becomes widely enough adopted
B) Enables features/graphics that would interest me without incurring a large framerate hit

When upgrading, I look for the best bang for my buck no matter which company offers it.

Originally posted by: chizow
He's trying to make it sound like pie-in-the-sky type stuff like DX10.1

I don't have to, it already is. DX10.1 hasn't made any meaningful inroads in current games and neither has PhysX. Only time will tell if this will change for either one. But until that happens, DX10.1 and PhysX are nothing but checkmark features to the majority of gamers out there.

 

[chizow]

Originally posted by: Creig
Maybe you should leave the personal comments out of your posts and stick to the thread topic. FYI, my last card was an XFX G92 8800GT.

Was this before or after your 3870 and 4850 purchases? I'd assume it was after given the lack of perspective in your comments about the importance of RV770 relative to G92 in various threads, particularly the recent "most important video cards" thread.

I don't have to, it already is. DX10.1 hasn't made any meaningful inroads in current games and neither has PhysX. Only time will tell if this will change for either one. But until that happens, DX10.1 and PhysX are nothing but checkmark features to the majority of gamers out there.

It has already changed, as the latest Mirror's Edge and Cryostasis gameplay footage show. You'd have to be incredibly ignorant or incredibly insulated to ignore the trailers for production games that will release in the next few months. You can claim "we'll see" ad infinitum but the reality of it is PhysX is being adopted faster and is having a greater impact on games than DX10.1 will ever have. That's not pie-in-the-sky, that's tangible benefits.

 

[error8]

Actually both Physx and directX 10.1 both suck at the moment. There is only the Mirrors Edge that fully supports PhysX and adds something there, like some cloth and shattering glass effect ( uuuu) . Then we only have a couple of Physx Unreal Tournament maps, which are not that amazing really ( played those on my ex 8800GT ) and we have a couple of demo uninteresting stuff.

ATi has Stalker DX 10.1 title ( with the latest patch ) and again a couple of demos. If you ask me, none of these are important. But it seems like Nvidia has the upper hand for the future, since it already made some contracts with a couple of game developers, to make them adopt PhysX in future games. I haven't heard anything from ATi's directX 10.1 implementation in future titles, yet. We'll have to see.

 

[chizow]

Originally posted by: error8
Actually both Physx and directX 10.1 both suck at the moment. There is only the Mirrors Edge that fully supports PhysX and adds something there, like some cloth and shattering glass effect ( uuuu) . Then we only have a couple of Physx Unreal Tournament maps, which are not that amazing really ( played those on my ex 8800GT ) and we have a couple of demo uninteresting stuff.

ATi has Stalker DX 10.1 title ( with the latest patch ) and again a couple of demos. If you ask me, none of these are important. But it seems like Nvidia has the upper hand for the future, since it already made some contracts with a couple of game developers, to make them adopt PhysX in future games. I haven't heard anything from ATi's directX 10.1 implementation in future titles, yet. We'll have to see.

Cryostasis PhysX Demo
Its more than just Mirrors Edge. Seriously watch this and then come back and say PhysX sucks. You haven't seen effects like that in any games because they simply do not exist. Crysis is the closest in terms of water behavior, and its not even close.

Edit: A playable demo is due this week actually and the game is set to release Feb 9th.

 

[nosfe]

problem with PhysX supporters is that they seem to think that PhysX is the only thing that can do physics(and don't give me the "i'm not one of them" speech, every rule has its exceptions). The problem with PhysX is that it's nvidia only, new nvidia video cards only that is so developers won't bother with it unless they go with the way its meant to be p(l)ayed. DirectX10.1 is in the same boat, worst actually, because we all know that it'll die out the moment DirectX11 arrives while PhysX might survive.

I'm not saying that those vids with what PhysX can do are crap, they're quite good, especially that cryostasis one, but the thing is, will i upgrade my video card for some special effects for a couple of games? no i won't and a lot of people won't either. Most people don't upgrade video cards that often and until there'll be a large enough percentage of cards that can do PhysX well(geforce 8200 and all those don't count and those with multiple cards also skewer the charts) there won't be a lot of titles that support it. In 2009 there'll be a lot more that in 2008 but that's easy to do considering the staggering amount of titles in 2008 with PhysX support. The real question is how it'll fare against what OpenCL/DirectX11 will have to offer. If they'll port PhysX to OpenCL then there's a big chance that ati will support it but untill that happens PhysX will remain a gimmick(more or less) with a big chance of dieing because of other solutions that'll be supported by both companies

 

[Creig]

Originally posted by: chizow
Was this before or after your 3870 and 4850 purchases? I'd assume it was after given the lack of perspective in your comments about the importance of RV770 relative to G92 in various threads, particularly the recent "most important video cards" thread.

Its nice to know that when somebody has an opinion differing from yours, you dismiss it as a "lack of perspective". It definitely gives me perspective into what type of person you are.

Originally posted by: chizow
You'd have to be incredibly ignorant or incredibly insulated to ignore the trailers for production games that will release in the next few months. You can claim "we'll see" ad infinitum but the reality of it is PhysX is being adopted faster and is having a greater impact on games than DX10.1 will ever have. That's not pie-in-the-sky, that's tangible benefits.

"games that will release in the next few months" are not currently a "tangible benefit". That's a possible future benefit. Otherwise one could say DX10.1 is a "tangible benefit" because it's likely that certain future titles will incorporate it.

If these proposed games are released with proposed meaningful PhysX content, then it will be a "tangible benefit". But not until that time. The same goes for DX10.1

 

[chizow]

Originally posted by: Creig
Its nice to know that when somebody has an opinion differing from yours, you dismiss it as a "lack of perspective". It definitely gives me perspective into what type of person you are.

Except opinion really has nothing to do with it. If you're going to ignore facts and claim the 4850 was more important than the 8800GT, you lack perspective.

If you actually bought a 3870 instead of an 8800GT when given the choice, you lack credibility. So when did you buy that 8800GT again?

"games that will release in the next few months" are not currently a "tangible benefit". That's a possible future benefit. Otherwise one could say DX10.1 is a "tangible benefit" because it's likely that certain future titles will incorporate it.

If these proposed games are released with proposed meaningful PhysX content, then it will be a "tangible benefit". But not until that time.

No its not a possible future benefit, the benefit is already realized but has not yet been made available. These are more than just tech demos, they're tangible benefits in productin titles.

Like I said though, you can claim "wait and see" forever and when Mirror's Edge is released, you'll claim its only 1 title, and then 2 titles, 3 etc.....