5870 Vantage result...

RavenXXX2 · 3 Oct 2009 at 18:53

amigafan2003 said:
I was comparing the GPU scores - mines nearly 4500 faster for the same card?

Yeah, your CPU is driving your GPU score up in the GPU bench as well as the CPU bench, I thought that would have been obvious.

straxusii said:
I set my 5870 at the same overclock as you have raven, I guess in P mode it is cpu limited, try extreme?

Max rez is 1680x1050 so cant do extreme.

straxusii · 3 Oct 2009 at 18:57

Extreme @ your clocks Raven, I would guess you will get similar as the GPU becomes the limiting factor

RavenXXX2 · 3 Oct 2009 at 19:12

Run the high test, my result below. What drivers and OS are you on, are your CCC settings at default. And how can you hit 16k GPU score whilst I'm on 12k?

straxusii · 3 Oct 2009 at 19:24

I'm using the drivers off the ATI site, 22nd I think. O/S is vista 64 bit

johny dolan · 3 Oct 2009 at 19:31

With my pc all stock clocks cpu is 2.93
vantage high settings
3dmark score H10736
gpu score 10796
cpu score 10409
vista 64 sp2
ccc all stock

RavenXXX2 · 3 Oct 2009 at 19:38

So is this a W7 performance issue? I had a GTX 260 before my 5870, maybe the Nvidia drivers are causing issues? they were uninstalled of course.

I know my CPU has not degraded in performance because I get the same CPU 3dm06 score I always did.

madindehead · 3 Oct 2009 at 19:38

This is my score using Performance setting. i7-920 @ 4GHz and a Zotac GTX 275 AMP! Edition

Using PhysX of course, because I can =)

Might try running it with PhysX off soon, and see what it comes out with

straxusii · 3 Oct 2009 at 19:39

RavenXXX2 said:
So is this a W7 performance issue? I had a GTX 260 before my 5870, maybe the Nvidia drivers are causing issues? they were uninstalled of course. I know my CPU has not degraded in performance because I get the same CPU 3dm06 score I always did.

I think you are just cpu limited in this test, going from performance to high our GPU scores come closer % wise. If you could run extreme I guess they would be very similar.

RavenXXX2 · 3 Oct 2009 at 19:40

madindehead said:
This is my score using Performance setting. i7-920 @ 4GHz and a Zotac GTX 275 AMP! Edition

You have physx enabled so not a fair comparison.

straxusii said:
I think you are just cpu limited in this test, going from performance to high our GPU scores come closer % wise. If you could run extreme I guess they would be very similar.

That does not explain your higher GPU score with a lower clocked CPU than me, I should have a higher score than you...no?

johny dolan · 3 Oct 2009 at 19:44

When i went from gtx 260 to a 4870x2 i had to reinstall the os to get the x2 to work as it should so maybe that's your problem.I dont get my win7 until oct 22nd i think it is so i cant check win7 scores until then.

johny dolan · 3 Oct 2009 at 19:47

straxusii said:
I think you are just cpu limited in this test, going from performance to high our GPU scores come closer % wise. If you could run extreme I guess they would be very similar.

But i score higher with my cpu at stock speeds of 2.9.His q6600 @ 3.7 will be much faster than mine at stock so he cant be cpu limited in comparison to mine.I know we would both score higher with i7 at 4 gig but something is not right with his score.

RavenXXX2 · 3 Oct 2009 at 19:52

If any of you guys could run the benchmarks in RE5 or/and SF4 at 1680x1050 and 4xAA all max settings, that would be great and if you have the games of course. I don't really want to install my OS again unless I'm 100% sure it is the issue.

straxusii · 3 Oct 2009 at 20:14

RavenXXX2 said:
That does not explain your higher GPU score with a lower clocked CPU than me, I should have a higher score than you...no?

No, bear in mind with a Q6600 vs Q9650 you have a kentsfield core whereas I have a yorkfield core which are significantly faster.

straxusii · 3 Oct 2009 at 20:15

RavenXXX2 said:
If any of you guys could run the benchmarks in RE5 or/and SF4 at 1680x1050 and 4xAA all max settings, that would be great and if you have the games of course. I don't really want to install my OS again unless I'm 100% sure it is the issue.

Don't have those but can do crysis/crysis warhead or GTA4 if you have any of those?

RavenXXX2 · 3 Oct 2009 at 20:22

Nah, don't have any of them, how about 3DMARK06? the SF4 bench is available as a separate DL. looking at reviews, your CPU has minimal performance over a similar clocked Kentsfield, makes sense as they are the same architecture just one is on a smaller process and has a bit more cache.

straxusii · 3 Oct 2009 at 20:29

I'll try the SF4 bench, I don't think you are right about kentsfield vs yorkfield, I guess it's not huge but it is definitely quicker: http://www.xbitlabs.com/images/cpu/core2extreme-qx9650/overall.png

RavenXXX2 · 3 Oct 2009 at 20:33

Looking at a few more reviews does reveal your CPU to have a bit more performance than I first noted, I may well opt for a Q9550, the Q6600 is a bit long in the tooth now. Look forward to your SF4 results.

http://www.techwarelabs.com/reviews/processors/Intel_Q9650_Quad_Core_Processor/index_8.shtml

straxusii · 3 Oct 2009 at 20:43

Here's 3dMark06, 900 1230 clock still on the 5870

16xAF 16xAA

RavenXXX2 · 3 Oct 2009 at 21:19

Well thats closer, I think your CPU is making the difference, although it could be OS related but I doubt it. Lets see if your CPU can hack it in real gaming
benchmarks..

kylew · 3 Oct 2009 at 21:31

I still don't get why people still use 3D Mark 06.

I got higher than that with my Q6600 at 3.96Ghz and my 3870X2.

http://service.futuremark.com/compare?3dm06=5711900

That should be enough to show you that it's not a benchmark that'll provide consistent numbers.