3DMark 2005 released!

The Batman said:
Beta ATI drivers. We're questioning their authenticity right now. They give a MASSIVE boost to the XTPE [and only a thrid of that boost to the XT] but they don't seem to give any appreciable performence increase in games like Far Cry [choose another DX9 heavy game].

Also FC 1.2 in SM2b crashes under the new drivers. This is because GI has been disabled. Not only that but it's no longer a selectable option. Which rasies the question, if GI is disabled HOW THE HOLY FUCK is 3dMark05 running under SM2b.
Click to expand...


Any driver that lets the X800XT series smoke the 6800GT's and Ultras in a synthetic bench, but DOES NOT beat them in most games, is veeeerrrrrrrrrry suspect in any book...regardless of whether ATI or Nvidia does it, they BOTH have been guilty of it many times in the past...
 
TheRapture said:
Any driver that lets the X800XT series smoke the 6800GT's and Ultras in a synthetic bench, but DOES NOT beat them in most games, is veeeerrrrrrrrrry suspect in any book...regardless of whether ATI or Nvidia does it, they BOTH have been guilty of it many times in the past...
Click to expand...


Very true! Just want to see who will start it off this round hehe
 
Geh, my aa and af were on.

Since I stoped playing CS I turned them on.
 
You know even if the hotfix turns out to be legit, if it doesn't really effect gaming it'll just go even more to show that synthetics are bunk. Blah, I always bench new drivers using the D3 timedemo and the Far Cry Benching utility anyways. The Marks are only good for testing my OC.
 
rancor said:
Its just not an AGP memory bug that was implimented. If it was only that one bug they fixed the AGP version of the xt would be right around the pci-e version, which is not the case :confused: and what ATi did might be a valid optimization who knows, but they stated they weren't going to do any synthetic benchmark specific optimizations, which would be false under this pretence.
Click to expand...
The systems used was not the same for PCIe and AGP (Of course). The PCIe system only had 512Mb mem and the AGP system 1Gb but I the difference shouldnt be that big, I agree with that.

About approved drivers, I don't see 66.70 here: http://www.futuremark.com/community/drivers/?approved

The Batman: I have know ide what you talk about, what has FarCry to do with 3DMark05?
The 66.70 drivers give a major boost to the 6800s according to the people over at my "home" forum at sweclockers (swedish forum).
 
Look at all the 6800 users clamoring for reasons their cards aren't winning, haha. People seem to have it so set in their heads that the nv40 could never lose to the r420 in any way, shape, or form. It's just 1 benchmark guys, you win some you lose some. We got D3, they won HL2, etc etc.

Whatever.

3DMark05 - 4422
 
3d05.JPG


everything stock

I guess that's a good score :rolleyes:
 
hmm well with my AthlonXP Mobile 2600+ (214x11.5) 2.5ghz, 1 gig PC3200 and perfect 9500softmod@ 9700 360/315 I got a score of 2,278 Not bad I say.

My Gigabyte 6800 should be here soon though :)
 
Leadtek 6800 gt @ 375/1050, A64 3400+ Clawhammer, 512 corsair xms, 80 gig WD 7200 8 MB, score of 4450. Yes I can go past ultra speeds, but I do not like overclocking (usually keep it stock at 350/1000) 350/1000 I get about 4000.
 
S_Z said:
The systems used was not the same for PCIe and AGP (Of course). The PCIe system only had 512Mb mem and the AGP system 1Gb but I the difference shouldnt be that big, I agree with that.

About approved drivers, I don't see 66.70 here: http://www.futuremark.com/community/drivers/?approved

The Batman: I have know ide what you talk about, what has FarCry to do with 3DMark05?
The 66.70 drivers give a major boost to the 6800s according to the people over at my "home" forum at sweclockers (swedish forum).
Click to expand...


Yep its a wierd jump, I would expect around 5000 for that bug fix and alittle higher then nV because of the amount of polygons they are using. System ram really has no influence on this, I'm almost postive they aren't going over 256 mb at any stage, if they are they are all the scores will be very close, even the ATi's x800 would be closing in on the xt pe.

ah ok was thinking about the 66.21's ;)
 
S_Z said:
The Batman: I have know ide what you talk about, what has FarCry to do with 3DMark05?
Click to expand...

Sigh. Geometry Instancing. Ever heard of it? It's disabled and non-selectable in the hotfix. This breaks SM2b compatabilty in FC 1.2 yet 3dMark05 runs SM2b just fine. Are you starting to see the picture?

And NO there WASN'T a massive performence increase in the 66.70s. All they did was give back the performence that was lost when moving from 65.76 to previous 66.xx drivers. Moving from the 65.xxs to the 66.xxs caused me to loose 12 FPS in D3 at 1600x1200 noAA. The reason I kept the 66.xxs is because they also caused an increase of 2 FPS in D3 at 1600x1200 4xAA [trading pure speed performence for eye candy performence].

People really need to read the fuck up, especially people with Nvidia cards. They spend 400 bucks on a card and don't bother being anal about their drivers [which from the X800 Pro > 6800U fiasco we should all be aware just how important drivers are]. If they're that kickback they should have stayed with the FX series.

Also, about 'approved' drivers. WHO. GIVES. A. FUCK. Seriously, this is [H]ardOCP, as in Hardcore, not Hardly packing a dick. WHQL, FutureMark Approved, Microsoft Recomended, GITG, TWIMTBP, runs best on Alienware/Intel P4, they can all blow me.
 
The Batman said:
Also, about 'approved' drivers. WHO. GIVES. A. FUCK. Seriously, this is [H]ardOCP, as in Hardcore, not Hardly packing a dick. WHQL, FutureMark Approved, Microsoft Recomended, GITG, TWIMTBP, runs best on Alienware/Intel P4, they can all blow me.
Click to expand...

Ahahaha. You are my hero. <3
 
The Batman said:
Quote:
Originally Posted by acrh2
Who said that the 8.07 results were inflated?
Click to expand...

I did.
Click to expand...

True. I have to agree with you on this one. Here you actually make a valid point, unlike in the rest of your reply. YOU ARE the only idiot calling these benchmarks inflated so far. While a man with half a brain might scratch his head in disbelief, he would not go screaming bloody murder like an arrogant son of a bitch that you are. Before you have any proof whatsoever, why don't you STFU and stop FUDding. That was mainly the reason why I posted my message. Your limited brain power was obviously insufficient to see it, so you went on the rampage of pointless arguing. Strike 1!

Quote:
As far as I understand this, ATI is claiming that these are correct results that one would get with 256M x800 level cards.
Click to expand...

Nvidia told me I could fly. I believe them because I own a GT. It makes PERFECT sense.
Click to expand...

My original comment was to, once again, point out that ATI believed that the previous results were false, not the new ones. This was just a way of interpretation what the company has said, necessary for later to make my own point of why this whole situation
might be legit.
And once again, you make absolutely no point in your reply and make an ass out of yourself. Lets see if you can go for 3.

Quote:
ATI is also claiming that the lower scores with older drivers were due to a bug which became mostly apparent in 3dmark05.
Click to expand...

Then explain the following post:
also my 3dmark05 score went up from 5294 to 5672 a huge boost for just a driver change.
Click to expand...

I'm sorry, here I just can't comprehend what your are trying to say. Does this mean that a driver bug would not let one's scores go from 5294 to 5672? Is that a law of physics? Am I totally off base here? I'll give you benefit of the doubt here. You are so far holding at 2 consecutive idiotisms.

The poster who wrote that has an XT [non-PE]. It's almost the same damned card as the PE yet it only saw 400 point increase from the "hotfix", whereas the XTPEs are seeing 1200 point increases [that's 3x the amount].
Click to expand...

Here you confused yourself by reading too much. I have to recommend for you to take breaks no less frequently than every 10 min or your brain may overheat. Just like it did here. Several web reviews have shown that x800 xt and x800 xt pe enjoy exactly the same gains going from 4.9 to 4.10+ catalyst.
But there's also more, buddy. Even if it were true, and x800 xt and pe were *different", what would that prove? In line of your arguments that ATI is cheating here, I have to understand your comment as you saying that ATI has cheated for just one card, the xt pe, while was unable or unwilling to do so for the regular xt. While this notion would classify your as paranoidal schizophrenic, I'm willing to go with an explanation where you simply didn't think through what you were trying to say. Strike 3!


Quote:
For one, I believe them. Two reasons:
1) We've always known that x800 line *simply had more raw processing power* than the 6800 line (with the exception for OGL games).
Click to expand...

Uh no JimBob, I'll assume by RAW processing power you mean high resolution no AA/AF [since AA/AF is more driver dependant and is also API independant]. The 6800U typically beats the XTPE at 1600x1200 noAA/AF, so it has more RAW processing power [and it does it at a 125mhz lower clock]. Also at 1024x768 noAA/AF we're going to be more CPU bound so your statement has no bearing on the conversation in the first place.
Click to expand...

This would have to be a multi-answer on my part. Lets just count how many times in row you are trying make yourself look dumb.
1) "Raw processing power" doesn't mean resolution or filtering. It means fillrates and pixel/vertex shader processing speed.
2) Typically, XTPE beats 6800U at 1600x1200 no AA/AF in almost every review in D3D games/benchmarks. I said previously, no OGL games, where the accepted theory is that the ATI ogl driver sucks (which is also directly related to xtpe having more processing power than 6800U).
3) The lower clock IS the reason why 6800U has less raw processing power.
4) Also at 1024x768 noAA/AF we are still going to compare two different gpu's with exactly the same CPU. And the raw processing power would definitely play a major (but not the only) role in what kind of difference we will see in the scores from two gpu's.


Quote:
2) I own an x800 xt pe.
Click to expand...

The stupidty of that statement pains me. I used to own an FX, that somehow made all of Nvidia's shady bilinear shit perfectly acceptable right? Jesus Christ man, come up with something better.
Click to expand...

You have to be retarded not to see that it was a joke. Or just have zero sense of humor. So far, the former seems to be more likely.

Quote:
So, until someone proves that ATI is cheating, or someone discovers that ATI is using some simple trick that is possible (but not yet implemented) on NV hardware -
things are exactly the way they should be - x800 xt pe is beating the pants off everyone.
Click to expand...

For some reason I feel compelled to bust out with some Disney music.

So until someone can explain the disparity of the performence increase between the XT and XTPE I'm going to firmly lodge this "hotfix" in Stalin's rectum.
Click to expand...

Sure, everyone is entitled to his opinion, but you seem to be missing my point here: stop FUDding, asshole, before you have any proof.
You need to make your points a little more clear. So far, you've been writing as if your are half brain dead. With time (and age), perhaps, you will learn to be more concise.
 
S_Z said:
Don't you guys read any other forums? This is what the hotfix does:

sirerics own words:
"The driver update was really just an AGP update. The PCIe X800's already scored the higher number. We released the 8.07, to fix a local memory utilization issue on AGP only cards. It's not a "3dmark05" specific driver. But 3dmark05 gets a benefit because it certainly taxes local memory more than most apps today. Any other app that requires lots of local memory will also benefit. "

And they are approved by futuremark: http://www.futuremark.com/community/drivers/?approved
Click to expand...



I don't give a damn who approves them, they improve the speed of a silly SYNTHETIC benchmark, based on a non-existant game engine, by 40% or more, and do NOTHING in REAL games...that my friend, would be called a driver cheat or "optimization" for a specific bench....blah.
 
Wow. An entire post of flames. I'd respond but you didn't make any points, just flames.

Someday the assholes of the world will come together and build a statue of you acne.

And you still didn't explain the SM2b and GI discrepancy. Not that it matters, clearly owning an XPTE has caused you to ascened past the plain of the rational. :rolleyes:
 
Well it may be weeks before I can find a decently fast server to dl from. I will not wait for a 25kbps stream to dl it.

Also where is the list for the top cards so far? I cant find it anywhere on the futurmark site. Lemme guess you have to register for the priveldge of visiting their site to see them?
 
-=[ 1226 ]=-

AMD 3200+
1GB Corsair XMS Extreme DDR400
5950 Ultra 500/1000 with 66.32 drivers
SB Audigy 2 ZS
Giga-Byte GA-7N400 Pro2 [F10 BIOS]

Few other things here and there...

That's a pretty crappy score for what I thought I was going to get. :) I personally think I should of gotten a better score.
 
2) Typically, XTPE beats 6800U at 1600x1200 no AA/AF in almost every review in D3D games/benchmarks. I said previously, no OGL games, where the accepted theory is that the ATI ogl driver sucks (which is also directly related to xtpe having more processing power than 6800U).
Click to expand...

Its varaiable, check the benchmarks again, depending on which dx game it is, the aa and af performance is variable on the x800 lines. Which leads to the fact they are using an application depend aa and af algos. And the precentage of the aa and af performance hits change, While on nV its almost always the same. This is very noticable in the same games in different levels also.
 
DermicSavage said:
The CPU tests render all the grahics through the CPU instead of the GPU...

It is really meant for systems that lack any sort of video card worth playing with... So, I dont understand why they are still keeping it around besides to just test how good the cpu is(which doesnt matter much imo with all the physics being calculated in the demos)
Click to expand...

Ahhh thank you very much my good man :D
 
TheRapture said:
I don't give a damn who approves them, they improve the speed of a silly SYNTHETIC benchmark, based on a non-existant game engine, by 40% or more, and do NOTHING in REAL games...that my friend, would be called a driver cheat or "optimization" for a specific bench....blah.
Click to expand...


Which ATi said they wouldn't do :p
 
Why couldnt people just post in relation to their impressions to 3dMark05 instead of turning this thread into yet another ATI sucks and cheats thread like almost all other video card threads?
 
The Batman said:
Sigh. Geometry Instancing. Ever heard of it? It's disabled and non-selectable in the hotfix. This breaks SM2b compatabilty in FC 1.2 yet 3dMark05 runs SM2b just fine. Are you starting to see the picture?

And NO there WASN'T a massive performence increase in the 66.70s. All they did was give back the performence that was lost when moving from 65.76 to previous 66.xx drivers. Moving from the 65.xxs to the 66.xxs caused me to loose 12 FPS in D3 at 1600x1200 noAA. The reason I kept the 66.xxs is because they also caused an increase of 2 FPS in D3 at 1600x1200 4xAA [trading pure speed performence for eye candy performence].

People really need to read the fuck up, especially people with Nvidia cards. They spend 400 bucks on a card and don't bother being anal about their drivers [which from the X800 Pro > 6800U fiasco we should all be aware just how important drivers are]. If they're that kickback they should have stayed with the FX series.

Also, about 'approved' drivers. WHO. GIVES. A. FUCK. Seriously, this is [H]ardOCP, as in Hardcore, not Hardly packing a dick. WHQL, FutureMark Approved, Microsoft Recomended, GITG, TWIMTBP, runs best on Alienware/Intel P4, they can all blow me.
Click to expand...
GI is not part of "SM2.0b". GI is a SM3.0 (VS3.0) feature, ATI use a "hack" to get it to work on thier cards. Anyway 3DMark05 does not even use GI.
 
Abit AN7
XP-M 2400
768MB PC3500
R9500 -> R9700

2729
http://service.futuremark.com/compare?3dm05=15476

It detected that I had 1GB mem but I only got 768MB (3x256MB). It also said my drivers were not "Futuremark Approved" although they're CAT 4.9. Might have something to do with W1z's soft mod.

W1z: any chance of softmodding the new 8.07 Beta CAT?
 
You know whats goddamn annoying? Going to the video card section and seeing "ATI VS NVIDIA" at every other thread. Leave it at 1 thread and stem off it goddamnit!
 
Djin said:
Here's some information from Nvidia.

http://www.nvidia.com/object/IO_16083.html

They state they've hit a score of 7229. And they beat ATI...

Yeah, with TWO video cards... come on Nvidia!!! Don't make a fool of yourselves again. :eek:
Click to expand...

Ahahaha, only 7200 with a grand in video cards alone? So lame. People are hitting past 6K with XTPE's, 1000 points for an extra $500 on your PC price, great idea!

Hahaha, yeah I know it'll be different in actual games, it's still funny :)
 
phaelinx said:
You know whats goddamn annoying? Going to the video card section and seeing "ATI VS NVIDIA" at every other thread. Leave it at 1 thread and stem off it goddamnit!
Click to expand...

If you read all the threads they all turn into ATI vs. Nvidia regardless of what the topic is.

Like I said i'll have to wait decades to get it dl'd at a decent speed.
 
AdvModDev said:
Abit AN7
XP-M 2400
768MB PC3500
R9500 -> R9700

2729
http://service.futuremark.com/compare?3dm05=15476

It detected that I had 1GB mem but I only got 768MB (3x256MB). It also said my drivers were not "Futuremark Approved" although they're CAT 4.9. Might have something to do with W1z's soft mod.

W1z: any chance of softmodding the new 8.07 Beta CAT?
Click to expand...

My machine is better, but he scores almost twice as high as I do. WTF Madonion!
 
The Batman said:
You know even if the hotfix turns out to be legit, if it doesn't really effect gaming it'll just go even more to show that synthetics are bunk. Blah, I always bench new drivers using the D3 timedemo and the Far Cry Benching utility anyways. The Marks are only good for testing my OC.
Click to expand...


That's the ticket, I test too with Far Cry mostly, and toss in some D3 for OGL goodness...nothing like a round of 1600x1200 4x AA and 8x AF with D3 to test your openGL settings :p

For the record, I play Far Cry with "very high" settings, 4x AA, 4x AF, at 1024x768 in both online and single player stuations...very smooth, very good looking.
 
I refuse to run 3DMark2005 until I get an explanation for the 41% performance increase in ATI's latest drivers. Until I see this I consider the '05 installation file a waste of bandwidth. A bug fix would not lead to such a huge performance increase. It's insane to say so.
 
You must log in or register to reply here.
Back
Top