4670 Crossfire Performance

They make noise, too.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Woolley
Posts: 8
Joined: Sat Apr 04, 2009 8:19 pm
Location: Australia

4670 Crossfire Performance

Post by Woolley » Sat Apr 04, 2009 9:51 pm

Xbitlabs have done a review comparing the 4670 in crossfire against other cards:

http://www.xbitlabs.com/articles/video/ ... firex.html

Their conclusion seems to be that using crossfire 4670s gives excellent results - better games perfomance than either the 4850 or 9800 GTX+ (as single cards) and far better power consumption.

You would never buy a 4670 if you are solely after gaming performance, but it looks like it could be a good all-round quiet setup.

Based on this it seems that a system with passively cooled 4670s would give about the best games performance for silent (passive) system. Has anyone tried this? Or are there better options out there?
Last edited by Woolley on Sun Apr 05, 2009 2:18 am, edited 1 time in total.

yensteel
Posts: 109
Joined: Mon Dec 22, 2008 5:58 am
Location: Hong Kong

Post by yensteel » Sun Apr 05, 2009 1:58 am

I think the 4850, or a 9600 GT SLI setup are good options as well. Both options can still allow for both custom and manufacturer fanless cooling.

Personally, I would buy the 4850.
Last edited by yensteel on Sun Apr 05, 2009 2:43 am, edited 1 time in total.

Woolley
Posts: 8
Joined: Sat Apr 04, 2009 8:19 pm
Location: Australia

Post by Woolley » Sun Apr 05, 2009 2:21 am

Oops - you're right - I've made a typo. Meant to write 4850, not the 4870. I've edited the post to correct this.

yensteel
Posts: 109
Joined: Mon Dec 22, 2008 5:58 am
Location: Hong Kong

Post by yensteel » Sun Apr 05, 2009 2:32 am

I'll update my previous post then.

axee
Posts: 170
Joined: Sat Nov 15, 2008 1:01 pm
Location: Slovenia

Post by axee » Sun Apr 12, 2009 2:19 pm

Two HD4670 with accelero S2's ?? Sounds like a good idea, but what would be easier to cool, one 4850 or two 4670?

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Sun Apr 12, 2009 9:04 pm

two 4670's by far. just case flow. I would imagine one fan blowing across them from the front, or a couple. Front air blowing in might be enough I would think. Have to be the right case though.

To me its easier to cool and control. WOuld I do this? I duno, ive thought about it. Cheap option for nice results. I wonder how they idle together?

Olle P
Posts: 711
Joined: Tue Nov 04, 2008 6:03 am
Location: Sweden

Post by Olle P » Wed Apr 15, 2009 1:54 pm

I think Xbitlabs test is a bit off the mark, because of their system setup.
What person in his right mind would go lengths to spend less than $200 on the graphics when the rest of the rig is all state of the art?
I think the very high end CPU and RAM in use skew the results.

To make the test worthwhile they should use much cheaper CPU, RAM and motherboard. And when it comes to go cheap with CrossFire vs single card, don't forget the added cost of having a CrossFire capable motherboard in the first place! (For example, motherboards with socket AM2+ can be had for $50 with single 16xPCIe (Asus M3A), but costs >$70 with CrossFire capability.)

Cheers
Olle
Last edited by Olle P on Thu Apr 16, 2009 3:11 am, edited 1 time in total.

K.Murx
Posts: 177
Joined: Tue Mar 17, 2009 10:26 am
Location: Germany

Post by K.Murx » Wed Apr 15, 2009 4:15 pm

Olle P wrote:I think Xbitlabs test is a bit off the mark, because of their system setup.
What person in his right mind would go lengths to spend less than $200 on the graphics when the rest of the rig is all state of the art?
Me :P
I do have uses for CPU and RAM to do actual work. The graphics cards are just there for the occasional game. So why spend $200+, when $150 will do?
Passively cooled 4850s are scarce compared to passive 4670s if you don't want to void your warranty. And the 4670s draw significantly less power (15W under full load, and a whopping 32 W less in idle - which they will be most of the time). Which is very nice. Also, the heat is spread out more evenly, which should make it easier to disperse with a fan, or am I wrong on that one?

Depending on where the power/performance of the 4770 falls, I might even get two of these for my next build instead of the one passive 4770 I thought about until reading that article.
I did not know that Crossfire scales that well - I thought up to now that the gains were more in the range of 50-70%.

If only somebody would make a (passive) 4670 X2... That would be nice. 4850 performance with much better power consumption. And the cost savings due to only one PCB should make it at least comparable in price...
*daydreams*
Last edited by K.Murx on Thu Apr 16, 2009 9:06 am, edited 1 time in total.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Wed Apr 15, 2009 8:10 pm

the two cards over the other options, 100%

Woolley
Posts: 8
Joined: Sat Apr 04, 2009 8:19 pm
Location: Australia

Post by Woolley » Thu Apr 16, 2009 5:42 am

Olle P wrote:I think Xbitlabs test is a bit off the mark, because of their system setup.
What person in his right mind would go lengths to spend less than $200 on the graphics when the rest of the rig is all state of the art?
Me too.
I bought a good CPU and plenty of RAM so I can do video editing and photo editing. I only play occassional games, so I don't want to spend an extra $150+ on a high-end card, particularly since most of the time it will just generate noise and heat.

I like the idea of a 4670 X2 card. That has started me daydreaming too!

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Thu Apr 16, 2009 6:32 am

yeah, and soon the 4670's will even drop in price.

xfire has great scaling in gaming. At ultra huge ends it shows up less, but at the 4670 end, it almost doubles performance.

Post Reply