• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Stress Test: ATI Tool vs. 3DMark06

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

golfking222

Member
Joined
Dec 9, 2006
Location
New York
What stresses a video card more, ATI Tool's little square or 3DMark06? My buddy here with his Crossfire 3870's is telling me 3DMark06, but I don't think it's worth the effort of watching 3DMark for 20 minutes as opposed to being able to tweak clocks with ATI Tool and watching for artifacts and such at the same time. I'm guessing they both probably eventually will put the video card to full load, but my friend doesn't think ATI Tool is capable of creating artifacts and a load temp.

Thoughts?
 
I have found that the ATI tool tends to be better to test OC's as it heats the gpu up a lot more than 3dmark does.

3dmark has the temp going up and down as its not always using 100% of the gpu, whereas the ati tool definitely maxes the gpu to 100% from what I have seen from temperatures.

With my X1900 I would use the ati tool to find my maximum OC with using the rotating image to heat up the gpu. Once I found a stable artifact free OC I would then loop 3dmark06 for a while.
 
How does rthdribl compare?

Ive used it and it gets the temps pretty close to those of ATI tool.

Between ati tool and 3dmark I would say ATI tool has the advantage of being highly GPU dependent, while 3dmark also stress tests the cpu which could take stress time away from the graphics card.
 
What about fur? It's the same concept with ATI tool, both rotating fur things but it's full screen. I got a few degreess higher on it than ATI tool, what about you?
 
For me, ATITool shows errors and artifacts before 3dmark does, so I end up with a more stable clock. Ill test out rthdribl and fur, and see what I think of them.
 
Fur has the ability to be both more intensive and less intensive than ATI tool I believe. Crank up the AA and res and watch your card chug along. But if you are strictly looking for artifacts I would say stick with ATI tool for the short run as it seems to do a better job imho.

That being said, running all three is never bad. ATItool for the short term, FUR for longer runs and 3Dmark for gaming stability and e-peen scores.
 
my 3dmark06 run at 1680x1050 with 8xAA runs perfectly alright (except a low score) but ATi Tool shows artifacts within the first few minutes. Is my OC stable/safe for gaming?
 
Last edited:
You could try assassin's creed, which locks up while 3dmark06 and fur could pass (both for 4 hours).
 
Don't ask me why, but ATI Tool artifacts at stock settings on my 4870/Vista x64. I see the artifacts, but the tool continues on and says "0 artifacts found." Not sure what that's all about.

I like to run Crysis in DX10 mode for a few hours to beat the hell out of my card, personally. It's definately game/bench stable. rthdribl has always been a good alternative, imo, but you have to watch it for artifacts, although the same can be said for other visual programs that don't detect them for you.
 
Definitely trust your eyes. The only time I use these "stability test" is to leave it on overnight or when I am busy.
 
I have some basic questions as I am now stress testing my overclock.

What is "true stablity" for a gfx card. Prime95 I like to run 24 hours but what tool and for how long? I ran the ATitool last night for 60 minutes on the test mode without it just saying no error for 60.00 minutes. Does this mean I am stable? I swear in 3dmarks i got a strange flash in the screen once or twice.

With less clocks and combos I wasn't able to get 3Dmark06 without crashing finally I got a stable mark06 and tested the ATitool. Previously with a less close I was getting missing pixels at this and that delta every 10-30 seconds, is that an artifact?

So how long should I run the test mode in ATitools to determine stablity?

If the 3Dmarks has screen pops (when something flashes out of synch with the picture) does this mean "unstable"?


Below are my scores from last nights testing with a headline of the setup. The basic really seems more dependent on CPU clock then anything else

9800Gx2
Q6600
790i Ultra
Kingston HyperX 9-9-9-27 1.9 (2x2gb)

UNclocked CPU / 177.19 drivers
3DMark Score 13663
SM 2.0 Score 5060
SM 3.0 Score 6742
CPU Score 3853

STOCK DRIVERS / 177.19 drivers / Q6600 clocked to 3.2
3DMark Score 17461
SM 2.0 Score 6851
SM 3.0 Score 8085
CPU Score 5110

GeForce ForceWare 177.89 XP 32-bit drivers / Q6600 clocked to 3.2

3DMark Score 17557
SM 2.0 Score 6860
SM 3.0 Score 8184
CPU Score 5104

Same as above with Oc'd gfx card
core/shader/mem
700/1750/1049

3DMark Score 17751
SM 2.0 Score 6716
SM 3.0 Score 8522
CPU Score 5123

750/1875/1075

3DMark Score 18047
SM 2.0 Score 6847
SM 3.0 Score 8716
CPU Score 5121

CPU oc'd to 3.375
700/1750/1053

3DMark Score 18549
SM 2.0 Score 7137
SM 3.0 Score 8767
CPU Score 5378

CPU oc'd to 3.375
750/1875/1110

3DMark Score 18604
SM 2.0 Score 7015
SM 3.0 Score 8973
CPU Score 5347

CPU oc'd to 3.4
726/1815/1145

3DMark Score 18851
SM 2.0 Score 7202
SM 3.0 Score 9007
CPU Score 5407
 
Back