• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

3DMark 2005 released!

Nyfrin said:
With a x800 PCI-E, I got 5,248 with the Cat 4.8's, and 5,545 with the Cat 4.9's + hotfix on the system in my sig.

http://service.futuremark.com/compare?3dm05=23339

If there's a 1,200-1,400 point 'cheat' in the hotfix, I sure didn't get it.

Edit: My sig isn't working. :(

P4 3.4 GHz (LGA 775)
Asus P5AD2 Premium (Intel 925x)
1GB Corsair DDR2 PC2-5400
Sapphire X800 XT PCI-E
2x WD 74GB Raptors, 2x WD 250GB 2500JD's
Click to expand...


What do you get without the hot fix?

Should somewhere around what the 4.8's got
 
To W1z

Could you please softmod the 8.07 Beta CATS? I got voltmods on my R9500 and I'd have to remove them and my VGA Silencer to do the hardmod. I'm sure there are many other softmod users that would like to see if these new Betas do indeed improve 3DM05 and, more importantly, D3. Please, W1z, at your earliest convenience.
 
AdvModDev said:
To W1z

Could you please softmod the 8.07 Beta CATS? I got voltmods on my R9500 and I'd have to remove them and my VGA Silencer to do the hardmod. I'm sure there are many other softmod users that would like to see if these new Betas do indeed improve 3DM05 and, more importantly, D3. Please, W1z, at your earliest convenience.
Click to expand...
Yes! I would second this request as well!
 
Yah, got 1766, while some other programs running the background. My video card isn't even that good... using Catalyst 4.9. Whoever got the 1000 with the FX5900 might wanna look into redoing that test.
 
AdvModDev said:
To W1z

Could you please softmod the 8.07 Beta CATS? I got voltmods on my R9500 and I'd have to remove them and my VGA Silencer to do the hardmod. I'm sure there are many other softmod users that would like to see if these new Betas do indeed improve 3DM05 and, more importantly, D3. Please, W1z, at your earliest convenience.
Click to expand...

first i'll look at those "miracle" drivers to see what they do to 3dmark .. if early/mid next week no new drivers are out i'll softmod them
 
Batman:

I'll play your game. Since you seem to be rather fond of that post you linked, let's see about looking at the whole thread. Here are the XT PE scores (listed with poster's name, if you'd like to quickly check to make sure I'm not fiddling with anything):

5788 - ThisMonsterLives
5888 - plink
5788 - truffle00
5712 - Mister E
5921 - xyber

The mean is 5819, with a standard deviation of 84. The website rancor gave with a score of 5817 seems to fall in fine with this data, but no, we shouldn't trust it. :rolleyes: This is information obtained solely from this forum, and I did leave Uncle out since he did not indicate whether or not his X800 Pro VIVO has been flashed, and his RAM is only at 555 MHz. His score was 5728, though, so it did fall into the correct range anyway.

Maxx, PureBooYah, and James Earl Ray were the only ones with scores above 6000, and their cards are all heavily overclocked (Maxx's overclock can be found on the second page).

I fail to see the relevance your score of 5500 in relation to stock XTs. You may be running a GT, but it's overclocked past Ultra speeds...what do you expect to get? If XT PEs are scoring ~5800 on average, I see no problem with XTs getting 5500. Furthermore, I see no problem with XT PEs receiving larger boosts than XTs because optimizations tend to cause percentage and not absolute increases. I understand you may have trouble with numbers larger than 10 (after all, you only have 10 fingers), so let me try to keep this simple. Let's say an XT PE scores a 6, and an XT scores a 5. New drivers are released that give a 10% improvement. The XT PE gains 0.6, while the XT only gains 0.5. Not only that, but the XT PE's improvement is 20% (uh oh, it's time to include your toes!) greater than the XT's! Oh my goodness, we have a genuine conspiracy here! :rolleyes:

Maybe it's time for you to find another XT user's boost, because there is no way they were getting 5100 points before the new drivers.

The point of my original post was to point out the ridiculous nature of your post. You're trying to unearth some grand conspiracy where ATI is now boosting the XT PE's performance more than the XT's, in addition to implementing improper optimizations. To what end? Who knows!, although I'm sure you'll conjure up something. Your problem, though, is that even your own data doesn't support your argument. Hell, according to you, the XTs outperform the XT PEs with the 4.9 drivers! How do you expect people to take you seriously? And now comes the worst part.

Instead of dropping this whole thing and admitting defeat, you're going to remain in the realm of the absurd and claim that my numbers are fabricated. Conveniently, you'll tell me that people on the forum in the thread YOU SENT ME TO don't know how to benchmark and you'll point to some thread on nvnews that does support your argument. You will then make some sly remarks, probably questioning my intelligence, capabilities of spelling and/or grammar, and sexuality, all while pretending to act hurt by what I've said. You'll probably even throw a smiley face in there at the end, so your message isn't entirely negative and attracts the attention of a mod. Finally, since I don't have the time to post 19 replies a day, when I tire of this thread and move on you'll end with something along the lines of "What? What now biatch? Where you is? You can't take me! KING KONG AIN'T GOT SHIT ON ME!" and you'll convince yourself of your superiority.

Also, if you're just playing devil's advocate, but always side with Nvidia, what exactly does that make Nvidia?
 
truffle00 said:
Maybe it's time for you to find another XT user's boost, because there is no way they were getting 5100 points before the new drivers.
Click to expand...

http://www.hardforum.com/showpost.php?p=1026679701&postcount=1

As Futuremark would say, OWND. Guy with a XT [not sure if it's stock or overclocked, do something constructive with your worthless life and ask him] is getting 5100 on the dot.

Now would you like to continue ranting or do you believe you've grown disenchanted with the taste of your foot?

Also, you know all this score disparity is serving to undermine your point.
 
The Batman said:
http://www.hardforum.com/showpost.php?p=1026679701&postcount=1

As Futuremark would say, OWND. Guy with a XT [not sure if it's stock or overclocked, do something constructive with your worthless life and ask him] is getting 5100 on the dot.

Now would you like to continue ranting or do you believe you've grown disenchanted with the taste of your foot?

Also, you know all this score disparity is serving to undermine your point.
Click to expand...

I already know what he's running at, do you? You've yet to show me a person on 4.9s getting 5100 on a stock XT, and I've yet to put my foot in my mouth. It's ok though, little buddy, keep trying. :rolleyes: You're like the one-legged dog that could...except you can't.

And what score disparity are you talking about? I'm pretty sure everyone knows that an overclocked card will score higher than a stock card.
 
truffle00 said:
I already know what he's running at, do you? You've yet to show me a person on 4.9s getting 5100 on a stock XT, and I've yet to put my foot in my mouth. It's ok though, little buddy, keep trying. :rolleyes: You're like the one-legged dog that could...except you can't.
Click to expand...


Maybe you need to overclock your brain. It's not performing to spec. So let me spell it out for you. This guy's got an OCed up the wazzu XT. YOU [as in truffle00 if you're having trouble with that] said there was "No Way" a STOCK XT could be getting 5100 on the old drivers. I just linked your condescending ass to a heavily OCed XT which is getting 5100 and you say it's irrelevant? It's TWICE as relevant. It's OCed genuis, it should be FASTER by your standards (well, then again this is YOU [truffle00 if your having trouble again] we're talking about, maybe I'm placing too much faith in you).

Face it, alot of people [including Steve of HardOCP] got over a 1000 point boost from these drivers. Alot of them didn't. Welcome to the world of synthetics. Maybe now your mind can begin grasping what Kyle, Brent, and all the others have been saying. Synthetics need to go.

Like I said, the disparity in these scores are hurting your case, not helping. Most people who came from the 66.32s to 66.70s got about the same point increase [if they had comparable systems of course], but people going from the 4.9s to the 4.9s + hotfix are getting all kinds of different scores. ATi said this was a memory allocation bug, so why isn't the bugfix effecting everybody? Explain THAT to me oh mighty zen-master.

fallguy said:
Dont worry, I caught it before you edited it. Way to resort to name calling, again.
Click to expand...

Either add something to the point of discussion or STFU. Truffle get's to rant and name-call yet your hypocritical ass doesn't say a word. I say something [then edit it almost immediately] and you're jumping down my throat. Seriously, you act like you sold your soul to get that GT. Well I for one am sick of your goddamned angst. Take your rainbow shooting ass to Christian Camp.
 
The Batman said:
Like I said, the disparity in these scores are hurting your case, not helping. Most people who from the 66.32s to 66.70s got about the same point increase, but people going from the 4.9s to the 4.9s + hotfix are getting all kinds of different scores. ATi said this was a memory allocation bug, so why isn't the bugfix effecting everybody? Explain THAT to me oh mighty zen-master.

Either add something to the point of discussion of STFU. Truffle get's to rant and name-call yet your hypocritical ass doesn't say a word. I say something [then edit it almost immediately] and you're jump down my throat.

Find a new hobby, this one's turning you into a nancy-boy.
Click to expand...

Because you cant follow the rules, you jump on me? Uh, ok. Speaking of hypocritical, you tell me to add something to the point of discussion, yet you are calling people idiots. Yeah, thats helping.

Because bugs dont always effect everyone? Some people have bugs in drivers, and some dont. Thats pretty common knowledge.
 
fallguy said:
Because bugs dont always effect everyone? Some people have bugs in drivers, and some dont. Thats pretty common knowledge.
Click to expand...

This was a MEMORY ALLOCATION BUG THAT WAS EFFECTING ALL 256MB CARDS. That is straight from ATI's mouth. That's pretty common knowledge too.

Next. :rolleyes:

And CleanSlate, I'll take that hug. :D
 
Obviously something isnt effecting all cards. Or else all scores would recieve the same percentage of improvement.

You cant even post without some childish :rolleyes: Do they make you feel manly?
 
Batman:

You don't make much sense, and I'm not sure even you know what you're saying in that first paragraph. I made the point that a stock XT should not net you 5100 points. It should be less...less, as in less than 4800 because if the XT PEs are getting 4800 points (5800 new - 1000 increase = 4800 points, no fuzzy math there I hope), and the XTs are clocked lower, the XTs should be scoring less than 4800 points. If the XTs are scoring less than 4800 points, and they then go up to 5500 points, that's a 700 point increase AT THE MINIMUM. Stock XT scores would be closer to 4600, giving you a 900 point increase; taking into account the percentage increase the drivers give, I fail to see problem. Who is getting wildly different results?

My problem is you're making something out of nothing to, as usual, serve your agenda of smearing ATI.
 
fallguy said:
Obviously something isnt effecting all cards. Or else all scores would recieve the same percentage of improvement.
Click to expand...

Ya it could be that. Or it could be that whatever optimizations are in the new hotfix [notice I didn't say cheat or hack or whathaveyou] isn't effecting everyone equally.

You cant even post without some childish :rolleyes: Do they make you feel manly?
Click to expand...

Actually, seeing you throw a hissy fit over the [now editied] 'idiot' blurb made me a feel a tad bit more manly. It almost word by word mirrored a scene from an old issue of the pre-crisis JLA [like hell if I can remember that far back]. In it Batman had made an offhand comment about one of the JLAs enemies when Superman turned to him and went "Bruce, Language!". What a Boy Scout.

So Fallguy, don't you have some kittens that need rescuing from some evil trees? :eek:
 
I didnt throw a hissy fit. I didnt even post about it until you edited it out. The rules are there for a reason. If everyone goes around calling names, and trying to insult someone, then there is very little actual discussion going on.

I dont see why you get all worked up, resort to name calling, and multiple :rolleyes:'s. :shrug:
 
Getting back on topic, http://service.futuremark.com/compare?3dm05=18706

That's a stock X800 Pro. It's getting a good 500 points more than my system [and no I'm trying to pump my chest, I just think somethings off when a stock Pro is soundly beating a GT at near UE clocks].

Would somebody please explain this to me. Really, you'll find that when the lot of you start making some sense I'm actually quite an agreeable and mild-mannered guy. Sorry if that last part of my sentence infringed on any of your copyrights fallguy. Please don't sue me, I'm poor. :D
 
The Batman said:
Well ya, you skipped about 15 drivers inbetween. 61.80, 62.20/22, 65.62/73/76, 66.00/02/10/31/32/51/60/70. I probabley missed a few from there as well. Nvidia's been a busy body. If each driver gave you an extra 75 points you'd be set [and I know I missed a few in there, I never bothered with the 62s myself and skipped right to the 65s from 61.77 so my knowledge of them is rather limited].
Click to expand...

AND unlike the ATi betas it seems to be in all 3d aps i just tryed the HL2 stress test picked up atlest 20FPS there have yet to check D3

Edit : just ran a time demo 1280x1024 high 4xaa 38.2FPS tthas at stock for the BFG 370/1000
 
Elios said:
AND unlike the ATi betas it seems to be in all 3d aps i just tryed the HL2 stress test picked up atlest 20FPS there have yet to check D3
Click to expand...

I could KISS you. THAT'S EXACTLY what I've been trying to say all this time. It's not even about the 'ethics' [as if that word exisits in corporate america/canada] of this whole situation. What good is a hotfix that doesn't do much for real world gaming? I haven't seen anyone come out and say and that this hotfix netted them another 5-10 FPS in Far Cry, Halo, [insert DX9 heavy game]. D3 doesn't amount to much as it's OGL, different API then 3DMark05. That and it has a hell of alot more improvement room right from the get go.

Hm....that gives me an idea. I'm gonna need someone with an ATI card [preferabley a stock XT/XTPE]. Truffle, you're nominated. Do me a favor and bench D3 at 1280x1024 4xAA Medium Quality. Then please do it again at High Quality. Then give it one more shot at Ultra High Quality. Use the hotfix as well but leave out the ATI AI shader replacement thingie. Not that I have anything against it, it's just lets not introduce another vairable to this experiment. Science is kinda picky about only having one vairabile per experiment.

Once you're done with that, rollback to the latest ATI certified drivers. Now perform the same tests. If what ATI said is true about the memory bug, we should notice a large disparity in the performance median between moving from Mediuem Quality [which uses 128mb] to High Quality [which uses 256mb]. Now I'm not talking about a pure frames difference, that could simply be from the hotfix drivers being faster and not directly relate to the bugfix. No, what I'm talking about is this:

If what ATI said is true, we SHOULD see a much LARGER performence increase when moving from High to Mediuem under the older drivers than we should see under moving from High to Mediuem under the beta drivers. If anyone doesn't get what I'm trying to say tell me and I'll explain it in [hopefully] better detail. That performence disparity should be around 10% [this is just a figure I extrapolated from the 'mean' of 3Dmark05 scores boots as Truffle pointed them out].

Well...somebody get cracking. It's a hell of alot more constructive than seeing the bunch of us butt heads over nothing.
 
The Batman said:
Getting back on topic, http://service.futuremark.com/compare?3dm05=18706

That's a stock X800 Pro. It's getting a good 500 points more than my system [and no I'm trying to pump my chest, I just think somethings off when a stock Pro is soundly beating a GT at near UE clocks].
Click to expand...

Its about 1500 more than my stock GT, with 61.77 drivers, and a stock 3400+. It looks to me that he has a 3400+ too, just with a 200Mhz overclock. Are the drivers hes using the 05 drivers... ? I can run my 3400+ at 2.4gig easily to make the CPU speed and run it again, might do it tomorrow. Does CPU speed matter in 05, more than it did in 03? I suppose the newest beta drivers would help my score, but I dont really care what it is. Last few times I tried one of the various beta drivers, I had a lot of problems with them, and Im pretty happy with the 61.77's.

Its pretty obvious the numbers are all over the place with 05. I cant think of any game where a stock X800 Pro beats a GT by 25% or more. The only explanation that even comes close to me, is that its a modded PE. In any event, the results of 05 doesnt look to replicate game scores between cards. At least in the limited scores Ive seen.
 
hmm i wonder what i get in FFXI now

Edit :FFXI is peged at 30 FPS lock like D3 almost all the time and i get ~25 at the lower Jeuno AH this is at 1280x1024 4xAF in the drivers

thats about 15 better then i was getting so it looks like it is all DX games
id check Farcry but i dont have a base line to go on

NOTE: this is all at BFG stock i havent clocked the card back up yet other then for 05
 
fallguy said:
Its about 1500 more than my stock GT, with 61.77 drivers, and a stock 3400+. It looks to me that he has a 3400+ too, just with a 200Mhz overclock. Are the drivers hes using the 05 drivers... ? I can run my 3400+ at 2.4gig easily to make the CPU speed and run it again, might do it tomorrow. Does CPU speed matter in 05, more than it did in 03? I suppose the newest beta drivers would help my score, but I dont really care what it is. Last few times I tried one of the various beta drivers, I had a lot of problems with them, and Im pretty happy with the 61.77's.

Its pretty obvious the numbers are all over the place with 05. I cant think of any game where a stock X800 Pro beats a GT by 25% or more. The only explanation that even comes close to me, is that its a modded PE. In any event, the results of 05 doesnt look to replicate game scores between cards. At least in the limited scores Ive seen.
Click to expand...

From my understanding the CPU doesn't effect 05 scores that much [I score higher then people with 3500+s OCed to 2.5] with my stock 3200+.

http://www.hardforum.com/showpost.php?p=1026678397&postcount=43

Incidently you're right, he's got a 3400 at almost 2.5. But the score isn't coming from the CPU, like I said I beat people with Oced GTs [the particular one I'm thinking of had his GT at 418/1150, sorry don't feel like post hunting right now] running almost identical CPUs.

And I don't think it's a modded Pro, he says Pro in his sig, and besides, I think 05 would pick up a modded Pro as an XTPE [since that's what the Bios would read].
 
Batman:

I don't know if you saw it, but I did post again on the previous page. Now, I don't expect a reply to every one of my posts but I did want to let you know that it's there, and hopefully better explains my position on XT vs. XT PE scores in 3dmark05.

I will run the D3 test, although since this is a new card I don't know if the overclock is stable yet. If it is not, I'll back it down and rerun all the tests, in which case the actual speed should not matter since we are only looking at the differences anyway. I will install the Doom 3 1.1 patch, but not use any other tweaks. I do not have .Net and so I don't have the Control Center; this should mean that my card will not be using AI with the beta drivers. I will leave all settings at default in my drivers, and let Doom3 control the AA settings.

For each resolution, I will start Doom 3 at that resolution and run through the demo 4 times, ignoring the first result (to allow everything to cache properly and not affect the scores). I will average the 3 remaining results to give me 1 number. As mentioned before, I will restart Doom 3 but not the computer between resolutions. This is in the interest of time, and should not significantly change the results. Naturally, I will reboot when switching the drivers.

If you have any issues with the design of my experiment let me know, and I will change the experiment accordingly.
 
so it looks like the 66.70's work for all the games DX games i got next on the list is 3Dmark01and 03
 
Elios said:
so it looks like the 66.70's work for all the games DX games i got next on the list is 3Dmark01and 03
Click to expand...

Yeah, the 66.70's are shaping up to be quite nice for benchmarking. Stay tuned in a day or two to see how they all stack up.....should be interesting to see results from earlier releases. :)
 
Halfway there, I decided to do 1600 as well as 1280, so this will take just a bit longer...
 
My 9800 pro can't run it, is there anything stating the rquirements? I keep getting this not enough video memory bullshit. Are there setting I need to set? HELP!!!!!
 
And here it is, the comparison between the 4.9s and 8.07 Betas in Doom 3 and 3dmark.

Test notes:
-Configuration is in my signature.
-Between runs of 3dmark the computer was restarted.
-Between runs of Doom 3 at a given detail level the game, but not the computer, was restarted.
-Anti-aliasing was set at 4X in the game settings.
-Catalyst Control Center was not installed for either driver.


Numbers are in the following format: 4.9 score/framerate / 8.07 score/framerate / Percent difference

3dmark01SE 21902 / 21957 / 0.25
3dmark03 11929 / 11976 / 0.39
3dmark05 4473 / 5788 / 22.72

Doom 3 - 1280x1024
Medium 48.6 / 53.3 / 8.82
High 46.0 / 51.7 / 11.03
Ultra 42.8 / 48.3 / 11.39

Doom 3 - 1600x1200
Medium 35.3 / 38.1 / 7.35
High 33.8 / 37.1 / 8.89
Ultra 31.9 / 34.9 / 8.60


Swing away, Batman.
 
Brent_Justice said:
i'm havin a blast watchin everyone argue over silly 3dmarks

not just in this forum, its on every forum
Click to expand...

Well I am tossing in the towel...I needed a format/reinstall anyways, so I figured now is the time to freshen up. So after getting XP patched to SP2, all drivers installed, I ran the 3D demo called "3DMark 2005"....and got exactly the same score. :eek: :p Even using the "new" 66.70 drivers...why oh why did I even think about caring wtf that silly bench gave me...my e-penis got in the way I guess :rolleyes: 4547 is the most so far...

OK, fuck it, Far Cry and Doom3 run beyooteeful and smooth. Fuck 3DM05...back to real games. :)
 
Im surprised that my M6805 laptop even ran it.

I managed to get a final score of 1072 with a cpu score of 3218 :)

Now all I need to do is upgrade my GF4 4600 in my main rig...
 
Did a little benching with the ati hotfix drivers, results are a bit ... odd.
My system is a 3ghz p4, gig o ram, x800pro @ stock.
Doom3

4.9s 1280x1024
MQ no AA - 51.5
MQ 4x AA - 25.3
HQ no AA - 34.9
HQ 4xAA - 24.2

Hotfix
MQ no AA - 40.7 (wtf?)
MQ 4xAA - 26.3
HQ no AA - 39.4
HQ 4xAA - 26.3

I noticed the ledge in the upper corner is slightly darker with the hotfix, everything else looked the same. Also when I had to run the timedemo through twice because the first time through (on either driver any setting) the demo would pause every couple seconds. ? The second time it always played fine.

3dmark05, x800pro now at 560/560
4.9s - 3996, GT1 14.4 GT2 11.6 GT3 24.4
Hotfix - 5107, GT1 22.6 GT2 14.6 GT3 26.0

The pics are from a x800xt at 1280x1024 w/ 4xaa. I could not see any difference (Hell, the pictures are the exact same size). I ran all the tests through twice and everything came back around the same number. Personally I don't like beta software on my box so I am back to the 4.9s, I didn't really play anything else when I had the hotfix on my system to see if there was a difference, although I doubt there would be. My personal observations from the 3dmark tests were that in GT1 when the camera paned out and showed the whole room there was a large increase in performance with the hotfix applied, when it was showing a close up of only 1 or 2 guys the fps was about the same. I could not see any difference in the other two tests, of course the viewings were several minutes apart.
 
You must log in or register to reply here.
Back
Top