PDA

View Full Version : This is not happening


Pages : [1] 2

Kruno
06-16-03, 11:06 AM
http://firingsquad.gamers.com/hardware/msi_geforce_fx5900-td128_review/page8.asp

http://firingsquad.gamers.com/hardware/msi_geforce_fx5900-td128_review/page9.asp

This can't be real can it?
A $1200-$1400 video card losing against a $500-$600 video card (prices in AUD).

This not only makes little sense but it practically throws out my idea that the NV3x architecture can be used in the future.

OT:
Of course, it can be used in the future as long as the architecture gets an overhaul.
I'm sure it will.

Back OT:
I thought the NV35 was faster than the R3xx series of cards under 4x AA and 8x AF?

WTF???
Am I a retard for thinking it is?

Can somone clean this mess up and explain to me what's going on?
I'm in shock and I'm going to the hospital to get treated.

jpeter
06-16-03, 11:09 AM
Interesting numbers for quake3 and SS:SE have been shown
by Brandon Bell at FS.



Writen by DaveBaumann in beyond3d
http://www.beyond3d.com/forum/viewtopic.php?t=6459
We've talked about using custom demos in reviews since this 3DMark stuff blew up. Brandon Bell over at FS has done just that with both Quake 3 and SS:SE - normally these reside in the playground of NVIDIA, but look at the order of the scores when a non-standard benchmark is used: http://firingsquad.gamers.com/hardware/msi_geforce_fx5900-td128_review/page8.asp http://firingsquad.gamers.com/hardware/msi_geforce_fx5900-td128_review/page9.asp Not exactly the orders we'd normally expect to see for these two benchmarks.

Kruno
06-16-03, 11:14 AM
http://www.nvnews.net/vbulletin/showthread.php?s=&threadid=13456

Can a mod please merge my thread with this one as it is basically about the same thing?

Thanks

--edit--Whoops, I think it deleted the wrong one. But yours came first, so yours it is. Heh, you people need to give me more stuff to do so I can actually figure out the admin tools :p --/edit--

The Baron
06-16-03, 11:16 AM
Yes, that was me above. :p

digitalwanderer
06-16-03, 11:21 AM
Those are some spooky numbers....and they seem pretty consistant and realistic like the guy actually put some effort into it and knew what he was doing a bit. (No offense intended Ben...well, maybe a little. ;) )

I'd like to read a more indepth analysis of this....
(Thanks for the link to B3D...but me brain is gonna hurt again. :( )

Hellbinder
06-16-03, 11:32 AM
Ill tell you whats more spooky...

Ok I am having some problems with some of the other numbers in this review like these

http://firingsquad.gamers.com/hardware/msi_geforce_fx5900-td128_review/page12.asp

I simply do not believe that the Nvidia card is really doing AA+AF on this game. There is just no freking way that the Fx is going to be 39 FPS faster than the 9800pro with AA+AF applied.

I repeat.. NO WAY.

I find a couple other oddities in the numbers as well. Some of it can be accounted for becuase of the Nvidia cards Core Speed advantage.

rellingsen
06-16-03, 11:41 AM
Sure points to certain IHV's optimizing for common benchmarks. It's just way too easy to do. Most sites just use the standard group of benchmarks like 3DMark's, Quake 3 , ShaderMark, etc. If you can just optimize for the 10 most common ones you can appear to sweep the competition. It just goes to show that web sites need to get away from using only the "standard" benchmarks. I think custom made benchmarks-time demos are the way to go and will limit the amount of optimizations like clipping planes etc. they can screw with.

Kruno
06-16-03, 11:44 AM
Originally posted by rellingsen
Sure points to certain IHV's optimizing for common benchmarks. It's just way too easy to do. Most sites just use the standard group of benchmarks like 3DMark's, Quake 3 , ShaderMark, etc. If you can just optimize for the 10 most common ones you can appear to sweep the competition. It just goes to show that web sites need to get away from using only the "standard" benchmarks. I think custom made benchmarks-time demos are the way to go and will limit the amount of optimizations like clipping planes etc. they can screw with.

I would like to see an auto-timedemo generator in games. :)

John Reynolds
06-16-03, 12:08 PM
Originally posted by K.I.L.E.R
I would like to see an auto-timedemo generator in games. :)

That rather defeats the purpose of a running a repeatable series of frames, doesn't it?

Kruno
06-16-03, 12:10 PM
Originally posted by John Reynolds
That rather defeats the purpose of a running a repeatable series of frames, doesn't it?

Yeh, well it also defeats preset clip planes along with many other "optimisations". :bleh:

The Baron
06-16-03, 12:11 PM
Originally posted by John Reynolds
That rather defeats the purpose of a running a repeatable series of frames, doesn't it?
Well, if there are, say, 16,777,216 possible paths, anything within 100,000 would be very, very similar. And of course, you should be able to seed the thing with a preset number in order to do IQ benchmarks and the like.

Hellbinder
06-16-03, 01:17 PM
to back up what i was saying about those nascar numbers. Look at extremetechs Nacar numbers... Vs the 5800U

http://www.extremetech.com/article2/0,3973,920348,00.asp

Edit here are the results for Nascar V.s the 5900U

http://www.extremetech.com/article2/0,3973,1075381,00.asp

There is just no freaking way that the 5900U is going to be that much faster. There is no doubt that AA+AF was disabled in the FS review..

Also the IL2 numbers are faster on the Radeon cards in this review.. Vs the 5900U

http://www.extremetech.com/article2/0,3973,1075334,00.asp

I knew there was somthing fishy about these specific numbers in the FS review. If you consider this, on top of the Q3 and SS results it paints a VERY different picture.

Hellbinder
06-16-03, 01:25 PM
I wonder if this is another Driver Bug...

something along the lines of

{IF Benchmark request AA+AF = YES Then Force AA+AF= OFF}

Is it a bug? or perhaps anew *performance enhancing Feature* in the new Dets...

(obviously I am not a CC coder.. consider it more of a Driver Flow chart thing..)

swanlee
06-16-03, 01:33 PM
What is it going to take for some of you to realize what Nvidia is doing? It's like a constant bait and switch with each new card release and every time the card never lives up to they hype and people instead of buying the fastest card at the time decided to wait for the new Nvidia release. and you people keep taking the bait from them. The Nvidia pr people must be laughing their butt off that you people keep passing on good cards you can actually buy while waiting on Nvidia's next "great wolrd beater" card.

From what we know now even if nvidia's new card swept the benchmarks by 1000 pts it will probablty turn out they cheated.

How long does it take for some of you to catch on that you are being swindled?

The Baron
06-16-03, 01:37 PM
...and right now, we're all waiting for the R350 refresh (ah, a refresh of a refresh--how stimulating), and then we're waiting for the R420 or the NV40.

Every company does this. It's called "hype."

Hellbinder
06-16-03, 01:38 PM
Woah there Swanlee...

The Gforce, GF2 ultra, GF3, and imo GF4 all basically kicked some serious ass. These bait and Switch tactics this extreme are realatively new side of Nvidia.

Hellbinder
06-16-03, 01:41 PM
...and right now, we're all waiting for the R350 refresh (ah, a refresh of a refresh--how stimulating), and then we're waiting for the R420 or the NV40.

Every company does this. It's called "hype."

um well the difference here is the 9700pro and 9800pro are both awesome cards. The rumored Refesh for the 9800pro is really more of a Bragging rights XT (Extreme performance) deal. Its not like you actually need it to have super buff performance. It is worth waiting for imo.. becuase in this case its a known Quantity. Being basically a super clocked 9800pro.

Uttar
06-16-03, 01:42 PM
Originally posted by Hellbinder
Woah there Swanlee...

The Gforce, GF2 ultra, GF3, and imo GF4 all basically kicked some serious ass. These bait and Switch techniues to this extreme are realatively new side of Nvidia.

Well, Reverend did post an e-mail of Scott of 3DFX seriously wondering if some detonators in mid 2000 really increased the GF2 GTS performance THAT much.
Heck, for all we know, we might discover soon that the Voodoo 5 was really good compared to the GF2 GTS and nVidia resisted all these years through policies of cheating.

I doubt that's true. I really don't think it is, in fact. But that they never cheated before, I don't think so. Just not much.


Uttar

Hellbinder
06-16-03, 01:49 PM
I doubt that's true. I really don't think it is, in fact. But that they never cheated before, I don't think so. Just not much.

well on this we part.. Because dispite the fact that i think the above products were awesome.. I also feel pretty sure that Nvidia has a looooooong history of Cheating by inflating benchmarks. I have been saying to whomever would liten to me for over a year that I was SURE that Nvidia was inflating their Serious Sam, Quake 3, RTCW, JK-II scores. And today i am beginign to be proven right.

I think if you look closely as far back as the TNT2 they had started fiddling with well known Benchamrks to inflate scores. There is even some pictoral Evidence that they lowered IQ in Q3 accross teh baord to beat the origional Radeon. Just look at the [H] Radeon review and you can clearly see its a blurry mess compared to the Radeon.

Having said all that however does not take away the fact that the cards were nice in and of themselves. which is somewaht different from their latest cards imo, which really makes it more like bait and switch. It is just a shame that Nvidia imo Continually pulls stunts like this so they can win all the Reviewers benchmarks. I guess they built the Speed is King bed. Now they have to sleep in it.

Ninja Prime
06-16-03, 01:56 PM
IMO, Nvidia is asking people to run "Best Performance" quality when benchmarking, compared to ATI's normal, or maybe even quality settings. That would probably make the difference you're seeing.

Eymar
06-16-03, 02:15 PM
Originally posted by Ninja Prime
IMO, Nvidia is asking people to run "Best Performance" quality when benchmarking, compared to ATI's normal, or maybe even quality settings. That would probably make the difference you're seeing.

That or Nvidia drivers render at a lower quality at the same app settings compared to ATI's. I'm pretty sure most reviewers force IQ settings by the driver rather than using the app setting though. If not then this could causing the discrepancy in benchmark scores.

John Reynolds
06-16-03, 02:40 PM
Originally posted by Uttar
Well, Reverend did post an e-mail of Scott of 3DFX seriously wondering if some detonators in mid 2000 really increased the GF2 GTS performance THAT much.
Heck, for all we know, we might discover soon that the Voodoo 5 was really good compared to the GF2 GTS and nVidia resisted all these years through policies of cheating.

I doubt that's true. I really don't think it is, in fact. But that they never cheated before, I don't think so. Just not much.


Uttar

I probably shouldn't write this since it's hearsay, but when Nvidia bought 3dfx and the latter's software guys started working at Nvidia, the one common statement out of them was something along the lines of: "We suspected NV had been cheating. But we had no idea it was this bad!!"

Solomon
06-16-03, 02:52 PM
Originally posted by swanlee
What is it going to take for some of you to realize what Nvidia is doing? It's like a constant bait and switch with each new card release and every time the card never lives up to they hype and people instead of buying the fastest card at the time decided to wait for the new Nvidia release. and you people keep taking the bait from them. The Nvidia pr people must be laughing their butt off that you people keep passing on good cards you can actually buy while waiting on Nvidia's next "great wolrd beater" card.

From what we know now even if nvidia's new card swept the benchmarks by 1000 pts it will probablty turn out they cheated.

How long does it take for some of you to catch on that you are being swindled?

Amen Brother!!! :afro:

Regards,
D. Solomon Jr.
*********.com

Hellbinder
06-16-03, 03:01 PM
After talking with teh reviewer at FS (Brandon). He assures me that AA is on in nascar. He told me he updated the review with screenshots. check the above links for the update.

Editor's Note: We've received emails from a few of you questioning whether the GeForce FX cards are performing anti-aliasing in IL-2 Sturmovik or not. In light of the Splinter Cell AA issue, we felt it best to provide you with screenshots directly from IL-2 demonstrating AA in action on the GeForce FX 5900 Ultra reference board. As you can see, AA is working (4xAA with 8xAF, in one shot, just like our test conditions) but just barely for a "4x" setting. Please download the bitmaps to see for yourself.

So while AA is on it looks like they are stretching the limits of what *4x* AA is supposed to be... :rolleyes:

Still, i have having problems accepting the stock Nascar adn IL scores as well...

Moose
06-16-03, 03:12 PM
Speaking of bait and switch...

I'm very suprised that ATI hasn't renamed their AA settings to match Nvidia's quality. After all Nvidia has been playing bait and switch with AF quality settings for a while now. quality/performance/application etc..

ATI should change

2x to 4x
4x to 8x
6x to 10x or 12x

That way reviewers would benchmark at a more even Image quality setting.