3DMark11 Score

Warrior

[H]F Junkie
Joined
Oct 13, 2004
Messages
13,827
can anyone give scores of an 58xx card with/without a dedicated PhysX card?
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,905
Generation/platform of CPU seems to have a MASSIVE affect.
You sure about that?

Okay -so here's the test results for running my i7-920 + GTX460 rig at the slowest CPU clock speed I can. 1.6 Ghz. The video card is the Galaxy GTX460 Super O/C and I didn't mess with it to underclock or overclock for either test -- everything in the drivers default win7 x64 install.

I7-920 at 1.6Ghz with GTX460 and 12GB of RAM - BCC at 133mhz x 12 multiplier
http://3dmark.com/3dm11/85290 - resulting in 3547 score

I7-920 at 4.0Ghz with GTX460 and 12GB of RAM - BCC at 200mhz x20 multiplier
http://3dmark.com/3dm11/38694 - resulting in 3850 score

So at 2.4Ghz less I get within ~300 points of the same score on the same machine
Seems 3DMark 11 doesn't really care much at all...
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,905
about clock speed? or cpu generation/platform? You can't always equate the two :rolleyes:

You really think a Core i7 at 1.6GHz is faster than the 3GHz+ Core2Quads and Phenom II X4's flying around here? Really? :rolleyes:

If it's rating name/generation over the actual performance of the chip, than this is a worthless benchmark even for comparing against its own scores.
 

Red Falcon

[H]F Junkie
Joined
May 7, 2007
Messages
11,248
You really think a Core i7 at 1.6GHz is faster than the 3GHz+ Core2Quads and Phenom II X4's flying around here? Really? :rolleyes:

If it's rating name/generation over the actual performance of the chip, than this is a worthless benchmark even for comparing against its own scores.

I couldn't agree more.

CPUs are not clock for clock the same, unless they are from the exact generation and even then there are other variations involved.

Use CPUbenchmark, that's a great way to identify the true processing power of your CPU and see where it ranks among all the others. :cool:
 

cmay119

Limp Gawd
Joined
Nov 30, 2009
Messages
194
Depending on whether cmay run his card at stock or oc'd for that bench would be helpful to know. If he run stock I'd stand by my initial statement.

Here's a compare between my OC'd 460 and his 5870, and considering my card is at a 20%+ overclock his 5870 performs well (if it's at stock).



Here's a direct compare over at Anandtech - http://www.anandtech.com/bench/Product/162?vs=180

Bearing in mind also that at higher resolutions cmays 5870 would pull ahead even more.

I'm impressed with the physics score his similarly clocked i7-950 puts out compared to my Q6600. Obviously his i7-950x is newer and cost more but it's still impressive.

Sorry, I should've stated that I was running at stock speeds on the card. I haven't really had a need to OC the card, because I've been quite happy with the performance of it at stock settings. :)
 

TehQuick

2[H]4U
Joined
Jul 3, 2004
Messages
3,738
You really think a Core i7 at 1.6GHz is faster than the 3GHz+ Core2Quads and Phenom II X4's flying around here? Really? :rolleyes:

If it's rating name/generation over the actual performance of the chip, than this is a worthless benchmark even for comparing against its own scores.

First, benchmarks by definition are not 100% reliable. Why should this one be any different? Secondly, you really think that Core i7 at 1.6ghz is the same at 3ghz+ C2Q? Really? :rolleyes:
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,905
First, benchmarks by definition are not 100% reliable. Why should this one be any different?
You should be able to run the same benchmark on two systems and be able to tell which one scored better.

If it's just arbitrarily increasing your score because your CPU reports as an i7, it's worthless.

Secondly, you really think that Core i7 at 1.6ghz is the same at 3ghz+ C2Q? Really? :rolleyes:
No, and I never said they would be the same. The 1.6GHz Core i7 would be slower than the 3GHz+ Core2Quad in pretty much everything except memory bandwidth.

Regardless, the guy tested the same Core i7 at 4GHz and at 1.6GHz. He cut the clockspeed of that chip in half (and then some), so there should have been a substantial drop in points if the CPU had much stake in the score, but he lost relatively little.

This can only mean one of two things. Either the CPU has very little impact on your final score, OR 3DMark is arbitrarily adding points to systems that report as having an i7 CPU.
 

Plato

n00b
Joined
Oct 28, 2010
Messages
8
It doesn't read my clock speed correctly, it's actually 3.8 Ghz.

3dmarkq.png
 

Arthur Hucksake

Limp Gawd
Joined
Oct 27, 2010
Messages
373
Strange old benchmark.

Uses the wrong SLI mode for a start for 5xx users, scores thrown about by similar CPUs.

Quite strange. Really shows how dependant we are on drivers asmuch as hardware it seems.
 

phatbx133

Gawd
Joined
Oct 27, 2006
Messages
726
My score result with 2 x 450 GTS SLI with Nvidia 265.90 quadro driver mod inf will work SLI mode.


Small bug, 3dmark2011 detect one video card *yellow highlight*

Capturebnjh.jpg


Capturehg.jpg


second run with 940 core 2200 memory close break 4K

bbbb.jpg


ccccc.jpg
 
Last edited:

DogsofJune

2[H]4U
Joined
Nov 7, 2008
Messages
4,022
Yeah, that work around took me from 3855 to 6753 ish. It's still buggy and meaningless.
 

Bloefield

[H]ard|Gawd
Joined
Nov 29, 2001
Messages
1,483
Just for the heck of it I downclocked my ic-7 920 from 3.3ghz [3.465 Turbo ] to completly default settings [2.67ghz , 2.80ghz Turbo ] and my score went down only 72 points.
P3805 > P3733.

By the way I'm still on the stock Intel hsf henceforth my mild 24/7 oc of 3.3ghz [Turbo = 3.465ghz]

3dMark doesn't seem to recognize the Turbo

http://3dmark.com/3dm11/134689
 
Last edited:

Dantrax

Gawd
Joined
Feb 9, 2007
Messages
705
(hd5970) X2504 - P6705 with stock Gpu settings & exactly the same after raising Cpu from 3.201 to 3.872. Next I got X2892 - P7726 using msi afterburner to set my (hd5970) to 850 gpu - 1200 mem - at 1.11 vdc (so stock 5870 xfire speeds) with Cpu still at 3.872 Mhz. Cat 10.11 drivers
 

Dantrax

Gawd
Joined
Feb 9, 2007
Messages
705
X2504 - P6705 at Stock speeds & X2892 - P7726 at 5870 speeds. This test seems to be very Gpu specific because I OCed my Cpu from 3.2 to near 3.9 with the same stock speed results. Demos look better at OC speeds on Extreme test but still not as good as Performance test Demos looked.
 

Blacklash

[H]ard|Gawd
Joined
Jun 16, 2004
Messages
1,893
Ran it on my slowest computer. 6850@ 945|1150 = P3914. That's with a Ph II 940 and old DDR2 ram.

http://3dmark.com/3dm11/147617

My 470s are gone. They're getting replaced with something else from nVidia or AMD. Going to make up my mind when the 6970 hits.
 

DarkStarr

Limp Gawd
Joined
Oct 25, 2010
Messages
267
A PhysX card shouldn't matter since 3DMark11 doesn't use PhysX (at least last I had heard)

EDIT: Rossi~ beat me to it.
 

pengs

n00b
Joined
Nov 30, 2010
Messages
11
P4720
http://3dmark.com/3dm11/156934

(single) ASUS GTX 470 810/1620/1800 1.037v
PII 955@3.7, NB@2200
TX750w

The only thing that is a little low is the physics score @ 4500ish with the PII. With the CPU clocked at 3.4 my score is P4500, at 3.6, no it's 4650ish.
 
Last edited:

Smoke1991

Limp Gawd
Joined
Jul 17, 2004
Messages
159
Im pulling 3800's
E8400 @ 3.0 ghz
2x GTX 470 SLI

But I also notice that SLI isint kicking in. tried doing the "fixes" but ill wait for nvidia to resolve this correctly.
 
Top