NEW - 3DMark 11 Compilation

[EarthDog]

Nope

HD5870 Still beats GTX470 in 3DMark 11 :

http://en.expreview.com/2010/12/09/futuremark-3dmark-11-benchmark-review/12823.html/14

@rickss69 : I will never surrender

It does there at stock speeds, but it doesnt show here in these results. Closest one is your 400+ points away. Even with the 5870 over 1k core with his hex vs my quad (which doesnt appear to matter) clocked 300Mhz higher.

Oh well.

Looking forward cadeveca's memory results though even though he's a grumpy son of a gun!

 

[fullinfusion]

can i post a few Xtreme runs just for the hell of it?

 

[HD4870]

I'm at 1055/1330 on air...

When systems are different , everything is possible ..... to many variables indeed

now that I've had a few days with the bench, i think it's safe to say that he's got better system memory than you do, perhaps?


That's what I'm doing right now as I type this..exploring how memory affects things...

It would be great to explore memory impact with 3DMark 11 and I have a Vertex 2 50GB SSD that "may" impact things too

Lots of variables..drivers, installed sfotware, blaah, blah, blah...welcome to benchmarking. Those are the things that keep it interesting.

Exactly

I like to keep my system very clean with only the necessary stuff (small size of my SSD forces me to keep that trend too)

It does there at stock speeds, but it doesnt show here in these results. Closest one is 500+ points away. Even with the 5870 over 1k core with his hex vs my quad (which doesnt appear to matter) clocked 300Mhz higher.

We al know that Fermi arcticture gains huge performance when overclocked and this seems good although a deeper look would suggest problems with the GF100 design in the first place


Edit : opps! My System specs aren't updated

 

[EarthDog]

I dont see why an SSD would impact a GPU benchmark. It doesnt for 01, 03, 05, 06, Vantage... This isnt PCmark.

Agreed on the overclocks. But hell hitting 1k core + and it beating it out in mist games though, tells me its the tesselation, which ATI cards are noteably weaker in compared to nvidia is what sticks out in my mind.

 

[cadaveca]

HDD speed makes a difference in quite a few 3DMark products. RAID over single HDD can = 1000 points in 06, dependant on platform.

Tesselation is also itneresting, if you own the bench, you can adjust/turn it off.

When it comes to the physics benchmarks, Fermi's unified cache is going to have an impact as well.

 

[cbupdd]

I use an unlocked version of afterburner, search it at google. I used 1.35v on my bench, but it's unstable with atitool
More voltage doesn't help at all.

 

[fullinfusion]

HDD speed makes a difference in quite a few 3DMark products. RAID over single HDD can = 1000 points in 06, dependant on platform.

Tesselation is also itneresting, if you own the bench, you can adjust/turn it off.

When it comes to the physics benchmarks, Fermi's unified cache is going to have an impact as well.

Na you cant turn Tesselation off even if you own the copy. I tried in both Extreme and Performance mode. If you drop both Tesselation sliders to there lowest setting and run the program you end up with a 0 score. I like that feature, it keeps the cheaters away.

 

[HD4870]

Updated my system specs :

System Name: High End
Processor: Core i7 920@4.25Ghz 1.3v
Motherboard: Asus Rampage III Extreme
Cooling: Noctua NH-D14
Memory: 3x2GB OCZ DDR3 2000Mhz Blade 9-9-9-24-1N
Video Card: VisionTek Radeon HD5870@1000/1300 (5200 GDDR5!)
Hard Disk: OCZ Vertex 2 50GB + 3x1.0TB WD Black + 2.0TB WD Green + 1.0TB WD Green + 640GB WD Blue
Optical Drive: Benq 1640 / Pioneer 111L / LiteOn 20A4P / Samsung 203B / NEC Optiarc 7200S
CRT/LCD Model: Samsung 2494HS 24" 1920x1080p LCD Monitor
Case: Gigabyte Aurora 3D 570 Case
Sound Card: Creative X-Fi Xtreme Music
PSU: Antec TPQ-850

 

[HammerON]

Nope

HD5870 Still beats GTX470 in 3DMark 11 :

http://en.expreview.com/2010/12/09/futuremark-3dmark-11-benchmark-review/12823.html/14

@rickss69 : I will never surrender

Maybe in that link you provided, but not on this thread

Earthdog and myself are topping all of the 5870's

 

[AphexDreamer]

Is it dangerous that I have to use 1.4 volts to get 1027??

Max temp is like 75C...

 

[erocker]

Is it dangerous that I have to use 1.4 volts to get 1027??

Max temp is like 75C...

Sounds about right for your high-leakage Asus card.

 

[HD4870]

Maybe in that link you provided, but not on this thread

Earthdog and myself are topping all of the 5870's

Don't forget that your OC is 41.7% / 19.4% For GPU / Memory while the highest OC with HD5870 is 22.3% / 11.2%

The bottom line is : HD5870 at stock clocks beats GTX470 at stock clocks in 3DMark 11 , the story with Unigine Heaven is over

 

[fullinfusion]

Don't forget that your OC is 41.7% / 19.4% For GPU / Memory while the highest OC with HD5870 is 22.3% / 11.2%

The bottom line is : HD5870 at stock clocks beats GTX470 at stock clocks in 3DMark 11 , the story with Unigine Heaven is over

Thank you! Story is over!

 

[HammerON]

Don't forget that your OC is 41.7% / 19.4% For GPU / Memory while the highest OC with HD5870 is 22.3% / 11.2%

The bottom line is : HD5870 at stock clocks beats GTX470 at stock clocks in 3DMark 11 , the story with Unigine Heaven is over

Very true about the stock clocks; however benching isn't about using stock clocks. It is about pushing your system to see how far she will go (at least it is for me)
Plus - that is the nice thing about the GTX 470 is it's ability to overclock so well

 

[TRIPTEX_CAN]

 

[HD4870]

Very true about the stock clocks; however benching isn't about using stock clocks. It is about pushing your system to see how far she will go (at least it is for me)
Plus - that is the nice thing about the GTX 470 is it's ability to overclock so well

Agreed

I was only talking about Tessellation-hurting-HD5800-more-than-GTX400 debate

Hmmm I wonder why experience is different than yours. I have had raid0 setups with both mechanical and ssd's and it made zero difference across xp, vista, nor 7 in any of those 3d benches. None of those benchmarks even us that in its equation to figure out the scores so im struggling to see how it would make a difference at all. Can u enlighten us on how that works?


Like someone else said if u turn off tesselation u don't get a score or it shows you didn't run the bench with default settings. So that can't be true. Correct?

Since physx runs om the cpu in 11 again im struggling to see how its unified arch makes a difference. I can test a bit later to see my physx scores with it enabled and disabled.

If you change anything in the settings , it won't give you a score, this is a fact (even with Vantage)

PhysX Enabled and Disabled has no effect on the score whatsoever

3DMark 11 uses Bullet Physics API (which is purely CPU dependent), not Ageia/nVIDIA PhysX

 

[NdMk2o1o]

Make way for NdMk


NdMk2o1o|GTX 470|760/1702|P4723|PII x2 550 @ B50 x4 3.5Ghz

I know theres room for an 800mhz+ core run on this beauty

 

[EarthDog]

Agreed

I was only talking about Tessellation-hurting-HD5800-more-than-GTX400 debate



If you change anything in the settings , it won't give you a score, this is a fact (even with Vantage)

PhysX Enabled and Disabled has no effect on the score whatsoever

3DMark 11 uses Bullet Physics API (which is purely CPU dependent), not Ageia/nVIDIA PhysX

Great info! I guess that's that!

 

[NdMk2o1o]

Great info! I guess that's that!

Double post FTL what about my score man, surely that's more important than anything???

 

[EarthDog]

Posting and moving around on a droid2 at this site seems harder than most for some reason.

Nice score. Keep pushing. And make sure u set you nvcp to performance so it runs the bench as it wants!

 

[fullinfusion]

Posting and moving around on a droid2 at this site seems harder than most for some reason.

Nice score. Keep pushing. And make sure u set you nvcp to performance so it runs the bench as it wants!

It's a Droid, Droid suck lol! 1.8ghx processor! :shadedshu

 

[MetalRacer]

MetalRacer|Asus 5850|980/1340|P4894|i7 980x @ 4412MHz

 

[HammerON]

Wondered where you where you were at Metal

Make way for NdMk


NdMk2o1o|GTX 470|760/1702|P4723|PII x2 550 @ B50 x4 3.5Ghz

http://img.techpowerup.org/101210/3d11 run1.jpg

I know theres room for an 800mhz+ core run on this beauty

Nice score, but you shouldn't have to use that much voltage to reach 760. On the three EVGA GTX 470's I have owned they didn't require any voltage increase until about 790. I usually use 1.020 volts for 860/2000 and they are very stable.

 

[erocker]

Posts removed.

Okay, okay enough.

Stop.

Really, stop. This isn't the thread for this.

Get back to posting benchmarks.

MetalRacer|Asus 5850|980/1340|P4894|i7 980x @ 4412MHz

http://img.techpowerup.org/101210/Capture141.jpg

I'm under budget in the CPU department. Must... add... voltage!

 

[jlewis02]

MetalRacer|Asus 5850|980/1340|P4894|i7 980x @ 4412MHz

http://img.techpowerup.org/101210/Capture141.jpg

Dang you now im the 3rd 5850 good job tho.