• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Underhanded ATI

S

SilverBack

Golden Member
We're told that on page 31 of the document, Richard advises himself: ?Steer people away from flow control in ps3.0 because we expect it to hurt badly. [Also it?s the main extra feature on NV40 vs R420 so let?s discourage people from using it until R5xx shows up with decent performance..

"Firstly, we wonder which people? ATI wants to "steer away" and "discourage"? As the target audience for the presentation was game developers, then it's reasonable to assume that these people are the gaming development community - and if so, this is atrocious."

"The way this could come across is that that for its exclusive commercial gain, ATI wants to stifle game developers - and thus limit the end-user experience ? until, that is, ATI is in a position to capitalise finacially."

 
here are the goodies:


from Anandtech's coverage
Slide 21
Key Messages:
.ATI is on the forefront of driving the visual experience in many different business areas
.We have moved beyond just a PC graphics company with a 3D focus
.This hits many market segments: consumer, commercial, SOHO & enterprise - to name just a few
.Also mention - higher integrated chipset volume than NVIDIA, P4 license, imageon, xilleon etc
.Setting targets for what features to aim for with games for late 2004.
.Note that many developers treat all DX9 NV hardware as "low end DX9" hardware because that's the best way.
.But be careful not to trash NVIDIA too hard here - your audience will help you out later...
.Also game genres obviously affect this and can drawn in low end DX7 h ardware too

Slide 25
.Setting targets for what features to aim for with games for late 2004.
.Note that many developers treat all DX9 NV hardware as "low end DX9" hardware because that's the best way.
.But be careful not to trash NVIDIA too hard here - your audience will help you out later...
.Also game genres obviously affect this and can drawn in low end DX7 hardware too

Slide 27
.True only of ATI's hardware.
.NVIDIA's hardware has 2-3 times the shader perf available in FF mode. But then that kills alls the useful features.
.Hence treat all NV DX9 hardware as DX8... (See previous slide)

Slide 31
.It's generally agreed that SM3.0 requires a move to 32 bit floats (some still think it's not)
.But - The spec requires it...
.Blend etc will be on NV40 and on our own SM3.0 parts
.Steer people away from flow control in ps3.0 because we expect it to hurt badly. [Also it's the main extra feature on NV40 vs R420 so let's discourage people from using it until R5xx shows up with decent performance...]
Click to expand...


Hence treat all NV DX9 hardware as DX8...

what an asshole. i feel sorry for him once the NV40 comes out. :|
 
It's definately a thing to watch and see what happens.

Good thing is that I believe that Nvidia will take back the crown this time, giving ATI a little humility that apparently it really needs.
It can only help us as this will spark off more competition between the two. End user wins again 🙂
 
Warning: Defensive post ahead.

Way to skew the presentation, guys!

Silverback, last I checked, promoting your own hardware within the bounds of reality and truth isn't considered "underhanded." And why not read the whole article? Paul goes on to say, "Apparently in one of its own GDC briefings, Nvidia itself has suggested that code branching in Shaders is presently best avoided."

Also, this quote:

"The way this could come across is that that for its exclusive commercial gain, ATI wants to stifle game developers - and thus limit the end-user experience ? until, that is, ATI is in a position to capitalise finacially."

Seems awfully similar to what nVidia did at NV30's launch, no? They continue to do so by promoting PS1.4 over PS2.0, at least as far as I've read. I'm not sure either company is out of the woods WRT to humility in marketing, but ATi's product track record seems to allow them the benefit of the doubt, no? Why does ATi "need" humility? Are you confusing the company with some of its more ardent supporters (no doubt you'll include me as one after this post)? 😉

gf3pro, even nV advises ppl to use PS1.4 over PS2.0 with NV3x. And both Newell and JC have said NV3x runs badly with the standard DX9 paths. Take a look at newer game benchmarks sometime (e.g., Xbits' Winter roundup) to see how ATi is generally faster in newer games. And why do you assume NV40 will be fast with SM3.0? NV3x certainly didn't live up to its billing as PS2.0+.

People here dissed ATi's superior PS2.0 performance as demonstrated by 3DM03 and a bunch of other apps as irrelevant. Why is PS3.0 performance so important all of sudden, when we probably won't see SM3.0 games for a few months to a year, and the market probably won't have many decently performing SM3.0 cards out to support those games?

I'm all for competition, but let's keep the speculation in line with the facts, eh? Underhanded? IMO, no. Inadvertant? Probably, considering the dearth of real R420 info up to now. But misleading? We won't know whether Huddy was FUDdy or frank until we see NV40 and R420 benchmarks.

Edit: More newer game benchmarks. That was a great Inq article, BTW. Dave at B3D praises Paul as an intelligent reporter, and it shows in his analysis of this ATi doc. At least he didn't post a contradictory or confusing or outright incorrect fact, as Fuad does in just about every one of his reports. :\
 
Seems awfully similar to what nVidia did at NV30's launch, no? They continue to do so by promoting PS1.4 over PS2.0, at least as far as I've read. I'm not sure either company is out of the woods WRT to humility in marketing, but ATi's product track record seems to allow them the benefit of the doubt, no? Why does ATi "need" humility? Are you confusing the company with some of its more ardent supporters (no doubt you'll include me as one after this post)?

I have never heard this, the only thing I have heard Nvidia do is promote the use of FP16 instead of FP32 for their hardware and with good reason for games. PS 1.4 is not even an Nvidia designed spec but a tack on for ATIs R200. Why would Nvidia actively support a standard for their competitiors chips?

Besides the R300 ATIs product record has been nothing short of horrid. I think they require a couple of more wins before they can recieve the benefit of the doubt.

gf3pro, even nV advises ppl to use PS1.4 over PS2.0 with NV3x. And both Newell and JC have said NV3x runs badly with the standard DX9 paths. Take a look at newer game benchmarks sometime (e.g., Xbits' Winter roundup) to see how ATi is generally faster in newer games. And why do you assume NV40 will be fast with SM3.0? NV3x certainly didn't live up to its billing as PS2.0+.


Newell is a paid lackey and JC talks about custom paths for Doom3. Custom paths he has always wanted and Nvidia delievered. And Nvidia has never advised people to use PS 1.4 over 2.0. They just advise FP16 over FP32 for the NV3.x

People here dissed ATi's superior PS2.0 performance as demonstrated by 3DM03 and a bunch of other apps as irrelevant. Why is PS3.0 performance so important all of sudden, when we probably won't see SM3.0 games for a few months to a year, and the market probably won't have many decently performing SM3.0 cards out to support those games?


We dissed it because it was demonstrated on a benchmark that is utterly unrealistic. As for the games that actually use PS 2.0 somewhere in the game Halo is close, Tomb Raider apparently is hosed for all participants and all I hear from the ATI crowd is how bad Far Cry runs on their rigs at 1024X768 while it is smooth 98% of the time on my 5900 at 1600X1200. So PS 2.0 disparity is nothing like what 3dmark03 claimed and the lackey said either at shadey days.

As for the presentation it is not biggie but kind of funny they admit outright their GPUs are utterly worthless for flow control and hope they can keep developers at bay until they deliever a solution that is capable of doing this in the next generation.

As for the PS2.0\PS3.0 stuff. It wont matter for games for the next 18 months but it will be interesting to watch everybody defend the R420 series because it wont meet the standards for PS3.0 but people were so ready to jump on the NV3.x because it didnt have FP24.
 
You're wrong about both PS1.4 and JC's comments about the NV3x and the default DX9 paths (not a Doom 3-specific comment, as D3 doesn't even use DX), but I don't have links handy, and don't want to take the probably great amount of time it'd take to find them (sorry). (By your logic, can I call JC an nV lackey for D3 being TWIMTBP and probably a NV40 pack-in, not to mention JC's GDC comment that he has an NV40 but not an R420? Please.)

Look at my last link and at HOCP's latest review, and you'll see the 9600XT outperforming the 5700U in both Far Cry and Halo (and Painkiller) in real-world situations using FRAPS. I'll admit H's Halo graph numbers for the 5700U don't seem to match up with its table, though. Xbit's roundup also shows ATi's superiority in FC, H, and DE2. Newer, apparently more shader-heavy titles seem to favor ATi. The high end cards tends to reflect the mid-range's scores, and ATi catches up on some titles that seem to favor the 5700U over the bandwidth-deprived 9600XT (PoP:SoT, KoToR).

I suppose Huddy's dismissal of flow control is just words until we can see benchmarks prove whether NV40 can make use of it at a decent speed, but I've already quoted Paul quoting nV's advice to use flow control sparingly, as it's a performance killer.

I'll "defend" (or, more accurately, consider irrelevent) R420's lack of PS3.0 if it's faster than NV40 in PS2.0 and NV40 is slow in PS3.0. Otherwise, I'd have to give credit to nV for a tremendous correction of their NV30 "stumble." People didn't jump on NV30 for FP16 vs FP24, but because it was ass slow compared to ATi. These configuration minutiae (FP16/24/32, SM2.0/2.0+) may be interesting and relevent to developers, but us gamers should really only be interested in speed at comparable quality. I could care less if a 5900 used FP16 rather than a 9800's FP24, as long as it ran about as fast with about the same IQ.
 
You're wrong about both PS1.4 and JC's comments about the NV3x and the default DX9 paths (not a Doom 3-specific comment, as D3 doesn't even use DX), but I don't have links handy, and don't want to take the probably great amount of time it'd take to find them (sorry). (By your logic, can I call JC an nV lackey for D3 being TWIMTBP and probably a NV40 pack-in, not to mention JC's GDC comment that he has an NV40 but not an R420? Please.)


Dont think I am. I have never seen Nvidia ever tell developers to code PS 1.4 for the FX series of cards. They always tell developers to work with reduced precisions instead of FP32.

As for the JC remarks on DX9. All I remember is JC remarking on a standard compiler from Microsoft the R3.xx is typically faster. But this is standard as we all know the NV3.x requires more tedious coding instead of the brute force approach of the R3.xx

Sure you can call JC whatever you want. But until JC calls a press conference with Nvidia employees running custom demos nobody can touch or work on and makes marketing graphs displaying price to performance ratios. Then go and make personal insults at ATI for not having the custom extensions he wants or having crap stencil performance for his shadowing. I dont think the term lackey for JC will hold much water.

I'll "defend" (or, more accurately, consider irrelevent) R420's lack of PS3.0 if it's faster than NV40 in PS2.0 and NV40 is slow in PS3.0. Otherwise, I'd have to give credit to nV for a tremendous correction of their NV30 "stumble." People didn't jump on NV30 for FP16 vs FP24, but because it was ass slow compared to ATi. These configuration minutiae (FP16/24/32, SM2.0/2.0+) may be interesting and relevent to developers, but us gamers should really only be interested in speed at comparable quality. I could care less if a 5900 used FP16 rather than a 9800's FP24, as long as it ran about as fast with about the same IQ.


Yes they did and you know it. They also made a huge ass deal about Nvidia telling developers to try and use FP16 as much as possible.

As for the features. They are probably worthless. Just like the P2.0 is about worthless on the R3.xx and NV3.xx because neither of them are terribly fast at it. But people had no problem singing praises for one arch while dimissing the other. In the end both are about worthless except for developers to work on.

NV40\R420 will have some hardware that should make PS2.0 be a viable platform IMO.
 
I was being facetious in calling JC an nV lackey, and in doing so mocking your calling Newell an ATi lackey.

"Tedious coding" or simple DX9, recent reviews (like the three I showed you) prove ATi to be faster in newer games. You really can't argue with benchmarks, unless you think all three reviewers are mistaken. Their benchmarks also seem to validate Newell's disappointment in NV3x's DX9 performance.

People may have mocked NV3x, but it wasn't just for FP16's lower IQ (which, BTW, nV itself demonstrated upon the NV30's launch--look for their FP16 vs. FP32 shots), it was for slow FPwhatever performance. And let's not forget R300 was as fast or faster than NV30, but with better AA (which is much easier to spot in screenshots than nV's reportedly better texture filtering).
 
i think "underhanded" is not an inaccurate term.

they are also calling ps3 shader support a "paper" feature, stating it will be some time before anything makes use of it in games.. hmm.. didn't see them calling r300 DX9 support a "paper" feature.. it's marketing propoganda, and ati is using it now, just as nv has used it in the past. neither is really more "guilty" than the other. they've both done it countless times, and will continue doing it. such is the way it is, which makes this particular argument rather pointless.

gabe newell.... yea, sorry.. his comments make him seem like a "lackey", tho perhaps that term is a little strong. he's definately shown a lot more partiality (justified or not) than JC ever has. while his concerns may well be justified, he comes off more like an ati spokesperson than an independant developer, tho he's been noticeably out of the limelight since "shader day".
 
Originally posted by: Genx87

Dont think I am. I have never seen Nvidia ever tell developers to code PS 1.4 for the FX series of cards.
Click to expand...

Part of this is understanding that in many cases promoting PS 1.4 (DirectX 8) to PS 2.0 (DirectX 9) provides no image quality benefit.
Click to expand...

http://www.hardocp.com/article.html?art=NTIw


the fx cards take a quite a hit when runing ps2.0, even with pp hints. hence, for nvidia, ps1.4 is perferred when possable.
 
Yes and we all know that PS2.0 is the only feature in DX9
rolleye.gif
 
Last I read that document has been edited a few times and is not considered accurate. I wouldn't get too excited myself. :beer:
 
Originally posted by: SilverBack
We're told that on page 31 of the document, Richard advises himself: ?Steer people away from flow control in ps3.0 because we expect it to hurt badly. [Also it?s the main extra feature on NV40 vs R420 so let?s discourage people from using it until R5xx shows up with decent performance..

"Firstly, we wonder which people? ATI wants to "steer away" and "discourage"? As the target audience for the presentation was game developers, then it's reasonable to assume that these people are the gaming development community - and if so, this is atrocious."

"The way this could come across is that that for its exclusive commercial gain, ATI wants to stifle game developers - and thus limit the end-user experience ? until, that is, ATI is in a position to capitalise finacially."
Click to expand...

On one hand, I've been telling you guys this is just the way business is for years. Whenever nVidia has done things like this, mud is slung, righteous fury runs rampant, and you would think they were grinding up orphans to make hamburgers.

On the other, I haven't even read the other posts in this thread and I bet some ATI worshipers are trying to spin this already.
rolleye.gif


 
hold on, when the HL2 numbers came out, every nvidia fan was crying "who cares, the games isnt even out yet"

where the hell are games with PS3.0 support? what? there arent any? no? then why worry about it?
 
And let's not forget R300 was as fast or faster than NV30, but with better AA (which is much easier to spot in screenshots than nV's reportedly better texture filtering).
Click to expand...

Shut your yap Pete! Do not be dissing my nV30, I will hunt you down, lock you in a closet with a tape of me playing UT2004!

😉

BTW- I say this PS3 stuff is as irrelevant now as PS2 was at the launch of the 9700Pro. Ask me again in a year and maybe the answer will have changed. Same as it ever was, meaningless features to sell cards that won't be in games until the cards are obsolete.
 
Originally posted by: Rollo
Originally posted by: SilverBack
We're told that on page 31 of the document, Richard advises himself: ?Steer people away from flow control in ps3.0 because we expect it to hurt badly. [Also it?s the main extra feature on NV40 vs R420 so let?s discourage people from using it until R5xx shows up with decent performance..

"Firstly, we wonder which people? ATI wants to "steer away" and "discourage"? As the target audience for the presentation was game developers, then it's reasonable to assume that these people are the gaming development community - and if so, this is atrocious."

"The way this could come across is that that for its exclusive commercial gain, ATI wants to stifle game developers - and thus limit the end-user experience ? until, that is, ATI is in a position to capitalise finacially."
Click to expand...

On one hand, I've been telling you guys this is just the way business is for years. Whenever nVidia has done things like this, mud is slung, righteous fury runs rampant, and you would think they were grinding up orphans to make hamburgers.

On the other, I haven't even read the other posts in this thread and I bet some ATI worshipers are trying to spin this already.
rolleye.gif
Click to expand...

rollo, i have a 429 page pdf file the confirms nvidia grinds up orphans to make hamburgers to feed their techs 😉
 
I fail to see how an internal company presentation can be classed underhanded, especially since something that is really underhanded like nVidia's blatant cheating has been dismissed on the grounds of "it doesn't matter how they get the performance as long as they get it".

It's also quite amazing to see individuals worried about PS 3.0 performance even while dimissing ATi's current PS 2.0 advantage as irrelevant.

Hence treat all NV DX9 hardware as DX8...
Click to expand...
Developers have doing this for quite some time; the NV3x is quite slow at PS 2.0 in case you hadn't noticed.
 
Originally posted by: ronnn
Last I read that document has been edited a few times and is not considered accurate. I wouldn't get too excited myself. :beer:
Click to expand...

Yup, me too. Could be as authentic as the Nigeria/Iraq nuclear materials report.
 
Originally posted by: BFG10K
the NV3x is quite slow at PS 2.0 in case you hadn't noticed.
Click to expand...

that's like saying you'd rather eat 6 pieces of sh** rather than eating 7-8 pieces of sh**. (PS2.0 performance is terrible for what you get from it with the current cards)
 
NV and ATI both play dirty in marketing, the online community has just chosen to skewer NV alone for some reason. It's like politicians - they all have mud on their faces and skeletons in their closet, but which ones you see are only the ones you CHOOSE to see.

Ahem, that's to you fanbois, boys, or however the hell it's being spelled these days.

When the benchmarks come out, whoever is the better value will get my money. End of story.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top