Motherboards Memory Storage Cases/Cooling/PSUs IT Computing Displays Mobile Mac CPUs & Chipsets Video Digital Cameras Linux Gadgets Systems Trade Shows Guides Home Increase Font Size Decrease Font Size Change Page Size
AMD's CrossFireX: Tri & Quad GPU Preview
AMD's CrossFireX: Tri & Quad GPU Preview
Date: February 21st, 2008
Topic: Video Card
Manufacturer: ATI
Author: Anand Lal Shimpi
Buy the VisionTek 900209 Radeon 3870x2 ATI
Blank
 Buy.com $385.99
 eCost $405.99
 MacMall $411.00
 
 

Power Consumption

In our 3-way SLI review we saw power consumption figures close to 800W at the wall outlet, thankfully with cooler running GPUs the CrossFireX numbers aren't as bad:

Number of GPUs Idle Power Load Power (Bioshock)
1 x Radeon HD 3870 148W 257W
2 x Radeon HD 3870 (1 X2) 181W 361W
3 x Radeon HD 3870 (2 X2 + 1) 211W 406W
4 x Radeon HD 3870 (2 X2) 240W 538W


With four GPUs we're over 500W of power consumption at the wall when running our Bioshock benchmark. There's still no clear need for greater than 1kW power supplies, but the better-safe-than-sorry mentality appears to be in full effect.

Final Words

We have to give AMD credit, there was no cherry picking of titles for this preview - for the most part, the benchmarks AMD itself selected showed no real need for 4-way CrossFireX over 3-way. We do appreciate the honesty, but it's clear that the world just isn't ready for a quad-GPU solution.

Due to the state of AMD's driver optimizations DX10 games currently only scale well to 3 GPUs and not much beyond (Crysis/Bioshock), while DX9 games will generally scale better all the way up to 4 GPUs. We expected the opposite to be true but AMD provided us with technical insight as to why it is the case:

"The biggest issue is DX10 has a lot more opportunities for persistent resources (resources rendered or updated in one frame and then read in subsequent frames). In DX9 we only had to handle texture render targets, which we have a good handle on in the DX10 driver. In addition to texture render targets DX10 allows an application to render to IBs and VBs using stream out from the GS or as a traditional render target. An application can also update any resource with a copy blt operation, but in DX9 copy blt operations were restricted to offscreen plains and render targets. This additional flexibility makes it harder to maximize performance without impacting quality.

Another area that creates issues is constant buffers, which is new for DX10. Some applications update dynamic constant buffers every frame while other apps update them less frequently. So again we have to find the right balance that generally works for quality without impacting performance.

We are also seeing new software bottlenecks in DX10 that we continue to work through. These software bottlenecks are sometimes caused by interactions with the OS and the Vista driver model that did not exist for DX9, most likely due to the limited feature set. Software bottlenecks impact our multi-GPU performance more than single GPU and can be a contributing factor to limited scaling.

We’re continuing to push hard to find the right solution to each challenge and boost performance and scalability wherever we can. As you can see, there are a lot of things that factor in."

From AMD's explanation it sounds like there's still a lot of work to be done on the CrossFireX driver. While we can expect to see its public debut in March, it seems like it'll be a while before we're anywhere close to ideal scaling. We've found ourselves in this position with many-GPU designs in the past, at least the players are taking things a bit more seriously this time around.

Home

 
  Index

Tools Share
Find lowest prices Find the lowest prices
Digg   del.icio.us   E-mail  
Print This Article Print this article  

28 Comments - Last by Anonymous Freak, 714 days ago
Username:
Password:
percents by skiboysteve, 719 days ago
"What were 70 - 100% gains from 1-to-2 GPUs are more than cut in half, becoming 30 - 45% gains when you add a third GPU"

this should be expected...

If each card provides 50fps in an ideal case...
1card at 50 fps
2card at 100fps = 100% improvement over 1 card
3card at 150fps = 50% improvement over 2 cards
4card at 200fps = 33% improvement over 3 cards

The most it can possibly provide is 100%, 50%, 33% improvements at 2,3,4 cards.

Noting that, COD4 scales incredibly well.



Reply
RE: percents by dustwalker13, 719 days ago
exactly what i was thinking ... cod almost scales ideally right up to 4 gpu's ... that 'somewhat reasonable gains' statement would fit HL2, which is scaling with every gpu added but the benefit diminishes. cod is scaling almost optimal:

gpus - max scaling - actual scaling
2gpu - 100% - 98%
3gpu - 50% - 44%
4gpu - 33% - 29%



Reply
RE: percents by Bladen, 719 days ago
I'll say.

I am very surprised to see "unfinished" drivers and tech preforming almost optimally, albeit in only 1 test case.

I think AMD should heavily encourage Infinity Ward to licence out COD4's engine.

Reply
RE: percents by Griswold, 719 days ago
"I think AMD should heavily encourage Infinity Ward to licence out COD4's engine."

Oh please, no!

Reply
RE: percents by Samus, 719 days ago
Yea, frak that. COD4's engine doesn't have the feeling or capability of the source engine, which scales great (except for the 4th GPU I suppose.)

With more tweaking, the Source engine is most ideal. Obviously they shouldn't target Crytek's engine, though.

Reply
RE: percents by RamarC, 719 days ago
i'd like to see how it handles mismatched cards in 3/4 way configs. another site did a crossfire review with a hd 3870 paired with a hd 3850 and their combined performance approached a hd 3870x2.

Reply
Amdahl's law ? by yyrkoon, 719 days ago
"From AMD's explanation it sounds like there's still a lot of work to be done on the CrossFireX driver. While we can expect to see its public debut in March, it seems like it'll be a while before we're anywhere close to ideal scaling. We've found ourselves in this position with many-GPU designs in the past, at least the players are taking things a bit more seriously this time around."

You guys, and many others have been talking about how well quad core CPUs scale(or dont) in a none server, are you really all that surprised to find that quad GPUs would be similar ?

I have been long suspecting that while the PCIe specification is absolute in how much bandwidth a number of PCIe lanes are supposed to have total, that motherboard manufacturers have been skimping as to how much they are allowing those lanes use. Maybe this is just now biting them in the behind ? Or maybe I am just being silly . . . ; ) Either way, system details are not exactly forthcoming here.



Reply
RE: Amdahl's law ? by skiboysteve, 719 days ago
... What?

Reply
RE: Amdahl's law ? by Pirks, 719 days ago
Do they speak English in "what"? (C) Pulp Fiction

Never mind, just kidding :)

Reply
RE: Amdahl's law ? by Griswold, 719 days ago
What!?

Reply
Comments Page 1 of 3

Deliver Rich Interactive Experiences
Engage more viewers with Adobe® Flash® Media Interactive Server 3.5
Vinpower DVD Duplicator
Manufacturer of quality disc duplication systems. Drop shipping available. Become a reseller today.
Unlicensed Software at Your Last Company
Anonymously Report Unlicensed Software with Our Form Now. Get Up to $1 Million.
Special Offer from The Economist
Get 12 issues of The Economist for $12. US subscribers only.
Unlicensed Software at Your Last Company
Anonymously Report Unlicensed Software with Our Form Now. Get Up to $1 Million.




Latest news by
DailyTech

 February 9, 2010

Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank
Blank

 February 8, 2010

Blank


more Video Card Discussions



pipeboost
Copyright © 1997-2010 AnandTech, Inc. All rights reserved. Terms, Conditions and Privacy Information.
Click Here for Advertising Information