3DMark05 Benchmark Results

[bkswaney]

This benchmark is just plain odd. The Pentium line is getting absolutely destroyed by the Athlon. I've never seen a benchmark where the CPU plays a HUGE role in overall score.

I have a Pentium Extreme Edition 3.4GHz @ 3.6GHz and I'm getting destroyed by a 2.4GHz Athlon! Hehe. :)

http://service.futuremark.com/compare?3dm05=16787
5270 marks

Regards,
Dan Solomon Jr.
*********.com


I agree 110% with that.
I'm using a Prescott clocked to 3.5 and getting eat up to. :(
Sounds like futuremark forgot to add our code or something. "lol" :screwy:

[boro]

I'm getting 5576 which I feel is pretty good for my rig :)
With 66.70

[schuey74]

5467 w/ X800 Pro @ 554/578

P4 3.0 @ 3405 (227 FSB)
P4C800-E DX
Gig of ram
8.07 Betas

http://service.futuremark.com/compare?3dm05=30565

[Blacklash]

Malficar, I get pretty much the same results with cpu at 1.6 or 3.2, so I agree with you. I'm benching to test stability and watch the purty slideshow, I haven't seen anyone here make too much of a contest, I'm actually suprised at how civilized this thread is! :) In any event, you illustrate very that a contest you be near pointless.

Understood. I wasn't interested in a contest, more of seeing how systems scale when they are running the bench at a baseline/default. You can't draw much information for score vs system if people bench with entirely different driver settings and application settings. The fact that there is such a glaring difference in the nVidia beta drivers and the 61.77's isn't helpful either. I was simply making a point with all the variables in a system ,scores are further skewed by not benching at a default value. I guess I believe these threads are good for sharing information/ system comparisons when there is a bench or measurement being used at a default value.

I appreciate it when people behave civilized and mature as well. Benches are interesting for testing. I pay more attention to measurements of dynamic game play than I do these tests.

I wouldn't suggest people ignore their CPU/Ram. Just that folks exaggerate its contribution/role in regard to other factors, particularly the video card.

Still 3dmark05 is very interesting bench and fun to watch. I think it is very well done.

[Warden13]

I agree with you guys that the A64 is destroying the P4 chips in this bench. I'm finding that a P4 with a similarly clocked GT has to be about 1.4ghz faster than my A64 in order to achieve similar results (saw a P4 at 4ghz score a tad higher than my A64 at 2563).

What seems odd is that people are saying overclocking their CPU gives little benefit. I think it's possible that CPU speed may not play a major role but perhaps the memory bandwidth and low latency of the A64 is playing a big role.

Warden

[Blacklash]

I agree with you guys that the A64 is destroying the P4 chips in this bench. I'm finding that a P4 with a similarly clocked GT has to be about 1.4ghz faster than my A64 in order to achieve similar results (saw a P4 at 4ghz score a tad higher than my A64 at 2563).

What seems odd is that people are saying overclocking their CPU gives little benefit. I think it's possible that CPU speed may not play a major role but perhaps the memory bandwidth and low latency of the A64 is playing a big role.

Warden
It's easy Warden. Drop your GT down to 425/1.10, leave your CPU where it is and bench. Then raise your video card back up to 460 and drop your CPU down to default, bench. Observe differences.

The last score I saw from JakUp was 5824, what AMD with a GT/Ultra at 460 or less is 'destroying' this?

[Warden13]

It's easy Warden. Drop your GT down to 425/1.10, leave your CPU where it is and bench. Then raise your video card back up to 460 and drop your CPU down to default, bench. Observe differences.

The last score I saw from JakUp was 5824, what AMD with a GT/Ultra at 460 or less is 'destroying' this?

Jakup's GPU is running 9mhz slower than mine but his memory is 46mhz faster than mine. His P4 is running with a 247 FSB vs 233 on my A64 and he is running nearly 1200 mhz faster tha my CPU (2568 vs 3715) yet I'm scoring nearly 300 points higher than him. In this bench, 300 points is significant, it's not like getting an extra 300 points in 2k01. I may be wrong since I haven't tested but I don't think 9mhz will add 300 points to a score, and again, he has me beat by 46mhz on his memory so he's getting higher bandwidth.

Warden

[Danhill]

Gave the 66.70 a run and got 5228

Come on Nvidia - give me some 3D mark 05 drivers

http://service.futuremark.com/compare?3dm05=34569

[jnd3]

The mighty machine clocks in with a ... 999 :retard:

[Blacklash]

66.70 Driver

Tweaks=No AA|AF set in driver panel, high performance driver settings.
No Tweaks=AA|AF application controlled, Quality driver settings.

CPU at flat 3.8
5544 440|1.15 tweaks
http://service.futuremark.com/compare?3dm05=34596
5382 425|1.10 tweaks
5246 425|1.10 no tweaks


Cpu at flat 3.5
5505 440|1.15 tweaks
5217 425|1.10 no tweaks

5544-5217= 327 for total score differences low to high.

Actual FPS differences between the best score and the lowest:

Game Tests
GT1 - Return To Proxycon
(fps) 22.4 vs 24.0


GT2 - Firefly Forest
(fps) 15.2 vs 16.1


GT3 - Canyon Flight
(fps) 26.7 vs 28.2


I have all of these saved and can only publish one at a time.

EDIT:
Final Note, Quality driver settings, AA|AF Application controlled, entire system at stock speed
= 5118.
CPU left at 3.0 and video card OC'ed with the "tweaks"
=5376.

Thus demonstrating how I can beat my legit 3.8 score with a 3.0c and "tweaking".

[Blacklash]

Jakup's GPU is running 9mhz slower than mine but his memory is 46mhz faster than mine. His P4 is running with a 247 FSB vs 233 on my A64 and he is running nearly 1200 mhz faster tha my CPU (2568 vs 3715) yet I'm scoring nearly 300 points higher than him. In this bench, 300 points is significant, it's not like getting an extra 300 points in 2k01. I may be wrong since I haven't tested but I don't think 9mhz will add 300 points to a score, and again, he has me beat by 46mhz on his memory so he's getting higher bandwidth.

Warden

You have a great rig. I would be interested in seeing your score with AA/AF application controlled and driver at quality settings, card frequencies 425/1.10 if it wouldn't trouble you.

The difference between my untweaked score at 3.8 and my tweaked score is 298points. Tweaks= AA/AF defeated, high performance driver setting, and 440/1.15 OC. Untweaked= AA/AF application controlled, quality driver settings, video card default frequencies.

[msxyz]

I've just downloaded 3DMark 05 free edition:

It scores an incredible... 703 points on my system :retard:

Sometime in the next weekend, I'm gonna try it with a Gainward FX5200 with 6nS RAM (DDR333) for a good laugh.
:lol2:

[Warden13]

You have a great rig. I would be interested in seeing your score with AA/AF application controlled and driver at quality settings, card frequencies 425/1.10 if it wouldn't trouble you.

AA/AF were app controlled on that run but I'll try and do a run at 425/1100 for you tonight.

Warden

[holmes]

Hopefully Nvidia will have some official drivers out before I hit retirement age.

lol

:lol:

[Blacklash]

AA/AF were app controlled on that run but I'll try and do a run at 425/1100 for you tonight.

Warden

Thank you for humoring me. Please be sure to use "quality" over high performance. I have a post a few up that shows my system with and without tweaks/OC on my video card. I can't help but be curious how bad I am beaten when our systems have the same video settings/frequencies.

[Arin]

3740
Seems low for me.
http://service.futuremark.com/compare?3dm05=8978

A64 939 3500+ @ 2288mhz 208x11
Gigabyte Nforce3 Mobo
2x512 3200 Corsair XMS Platinum 2-2-2-5
Evga 6800GT 351:1002

61.77 video
5.10 mobo

windows xp with sp2 and all updates.

I have a duel monitor setup, had the other monitor running. I left all the quality settings at "Highest quality", and "application prefference".

Should I expect more? What drivers should I try?

[Blacklash]

3740
Seems low for me.
http://service.futuremark.com/compare?3dm05=8978

A64 939 3500+ @ 2288mhz 208x11
Gigabyte Nforce3 Mobo
2x512 3200 Corsair XMS Platinum 2-2-2-5
Evga 6800GT 351:1002

61.77 video
5.10 mobo

windows xp with sp2 and all updates.

I have a duel monitor setup, had the other monitor running. I left all the quality settings at "Highest quality", and "application prefference".

Should I expect more? What drivers should I try?

66.70 get it from Guru3d< Try quality driver settings for your real score, IMO. If you want to inflate it, like many are doing, same driver(66.70), AA/AF off in your driver, and high performance in the driver. Some people are even inceasing LOD to high numbers, I haven't tried that. I gain 136points from those tweaks alone, and can gain up to 298 with those and a slight video card OC.(440/1.15 vs 425/1.10)

EDIT: for benching I'd disable that second display. I have seen people get terrible bench scores with more than one monitor designated, for whatever the reason.

[Arin]

If I remember correctly, I've tried a few driver upgrades from 61.77 and they always gave me BSOD's when it booted up...

Should I uninstall the drivers I have completely and then install the new drivers?

[Warden13]

Your best bet is to boot into safe mode, uninstall existing driver, run driver cleaner, reboot into safe mode, install new drivers, reboot and enjoy :)

Warden

[Warden13]

Oh yeah... I may be the only idiot on the boards but I don't find 05 to be that visually stunning. I mean the nature section is nice and all but aside from that I don't find it to be that "pretty". I might just be spiled by Doom 3 and, to a lesser extent, World of Warcraft.

Warden

[anzak]

Oh yeah... I may be the only idiot on the boards but I don't find 05 to be that visually stunning.

Return to Proxycon and Canyon Flight blew me out of the water. Looks much better than Doom 3 with it's low poly models IMO.

[nmyeti]

5609

evga 6800U at UE speeds

Not all that impressed with this score. I guess i can only hope for better drivers.
http://service.futuremark.com/compare?3dm05=39419

[Omega53]

nVidia needs to hurry up with new drivers :(

[Arin]

Jebus! The 6800GT will run that fast with just an NV5 and 1.4 mod?
I must 1.4 mod mine now.... i'll throw in the damn floppy drive.

[Omega53]

Jebus! The 6800GT will run that fast with just an NV5 and 1.4 mod?
I must 1.4 mod mine now.... i'll throw in the damn floppy drive.

Nah I can only get 440/1180 stable with the nv5