Are these results poor for an unlocked 6800xt?

madad2005

Hi,

I recently (today) bought and unlocked my xfx 6800xt. I managed to get 16 pixel shaders and 5 vertex shaders working (the final one was knackered). I tested at each step and got the following results (I'll show my previous result using the 6600 for show). I am wondering if these results are a bit low - I might have missed something:

6600:

Aquamark3: 33000; 3DMark03: ~5400; 3DMark05: ~1700

6800xt - stock settings (8ps,4vp@325/700):

Aquamark3: 45375; 3DMark03: 7387; 3DMark05: 2785

6800xt - 12ps,4vp:

Aquamark3: 50104; 3DMark03: 8779; 3DMark05: 3138

6800xt - 16ps,4vp:

Aquamark3: 52029; 3DMark03: 9417; 3DMark05: 3357

6800xt - 16ps,5vp:

Aquamark3: 53139; 3DMark03: 9689; 3DMark05: 3602

6800xt - 16ps,5vp@355/768 (using autotest in coolbits):

Aquamark3: 53750

the 12v rail on the PSU only supplies 17A, but I'd expect some instability if that was the problem, but it hasn't been at all. The rest of my pc specs are on the left. GPU temperature in 2D is at 60-63 deg celsius. I will reduce this when I add on my zalman vf700cu.

I have system performance set to quality. The rest of the options are set to application-controlled.

Any thoughts?

madad2005 · 01-10-2006, 12:57 | posts: 718 | Location: England

Ah well, doesn't really matter anyway. If the PSU is the problem, I'll be upgrading it in the summer. If the CPU or hard disk is the problem, it will have to wait till next January before it gets upgraded.

At the end of the day, this new card has boosted my performance tremendously compared to my 6600. I never knew games could look so good!!

Vinnie · 01-10-2006, 13:30 | posts: 7,290 | Location: The Netherlands

Iam not sure (never heard about the 6800xt), but the scores doesnt look good to me..

Altough, if the 6800xt is as powerfull as the 6800 then the scores r good enough

madad2005 · 01-10-2006, 13:57 | posts: 718 | Location: England

It is now roughly the same as a 6800, if not a bit better. Cheers for the response. This place gets pretty quiet at times.

eat@joes · 01-10-2006, 14:20 | posts: 1,014 | Location: Nottingham, MD

Nice job. The scores are good for the price of the XT. I tried to unlock the LE but had too many problems with it.

*ever reach for a cup on your desk and drink out of it, but realizing it's the wrong cup filled with old warm liquid, but instead of spitting it out you just drink it? I just did that

Jeremy · 01-11-2006, 08:05 | posts: 4,070 | Location: New Zealand

You should set the quality to performance and all Anti-Alising settings to off. That should boost your performance a bit.

jaiking · 01-12-2006, 06:07 | posts: 141 | Location: South Australia

then what the point of the benchmark if you cheat but disabling all the default settings??

yeah it pushes the score up but it doesn't really show what your card can really do.

dont mean to flame ya jeremy2223 if thats what it sounds like

btw with my bfg 6600gt oc before it died i was almost breaking the 50,000 mark so maybe play around with chipset and video drivers to see if they are holding you back any

eat@joes · 01-12-2006, 10:20 | posts: 1,014 | Location: Nottingham, MD

Hey Jaiking. The point of benchmarking as you suspected is definately to stress your card, but that's usually for self enduldgement and seeing how badly you can push your card. But, what we are trying to see over here at guru is how your card fares up against others. The default settings are with no AA and AF, and with image quality tuned up to performance. That way it gives you the "best possible" scenario. Then post settings with AA and AF and such after that to see the perfomance hit it takes (if any).

jaiking · 01-12-2006, 10:28 | posts: 141 | Location: South Australia

sorry my bad i actually forgot the default settings for aquamark as i dont have it on this system and seeing my 6600gt is dead i have no need for it,

i just ment thet with a lot of the benchmarks that have default AA and AF on people are are forcing them off to boost the scores, and them comparing them to people using default settings. i'm not saying that everyone does it but some do like most of my local LAN buddies.