Go Back   Hardware Canucks > HARDWARE > Video Cards

    
Reply
 
LinkBack Thread Tools Display Modes
  #1 (permalink)  
Old March 15, 2011, 04:08 AM
Hall Of Fame
F@H
 
Join Date: Dec 2010
Location: Fort St. John, BC
Posts: 1,222

My System Specs

Default SLI/CF testing? Best method?

Hey,

As I'm starting to look at benchmarks about multi-GPU performance (I have an itch to go surround, but I'm resisting for now), I begin to wonder how the best way to judge performance is?

Obviously, if I'm thinking surround, higher resolution, but after reading some pitiful (results) reviews on 3-way GPU benchmarks, I began to wonder if the fatal flaw of most of these reviews is not using a high enough resolution? Even if we run a i7-2600k @ 4.0 Ghz or a i7-750 @ 3.8 Ghz, is that enough to remove the CPU bottleneck @ 1920x1080, or even 2560x1600?

Will we see the best GPU scaling on a 3-way monitor set-up vs. SR-2 with 2 x i7-990x @ 4Ghz each @ 2560x1600?
Reply With Quote
  #2 (permalink)  
Old March 15, 2011, 04:14 AM
Arinoth's Avatar
Moderator
F@H
 
Join Date: May 2009
Location: Halifax
Posts: 8,598

My System Specs

Default

The bottleneck for multi-monitor I have found is still on the gpu side either with the overall bandwidth for pci-e slots or how fast the two can communicate and process images as the same time. I've run games on an e6400, i7 940 and a 2500k all at different clocks and i haven't noticed a notable difference in frames (or benchmarking results, though the e6400 did score significantly lower) between the i7 940 at 3.8ghz an the 2500k running stock.

You must remember that all newer platform, especially the 1366 is designed for high bandwidth for multiple pci-e devices, namely graphics cards, and really never sees a bottleneck at all. As for 1156 and 1155 i doubt eithe running at those speeds would cause a bottleneck even at 5760 x 1080, however i can only speak from personal experience and the minimal benchmarking i've done (typically just game and monitor fps)
Reply With Quote
  #3 (permalink)  
Old March 15, 2011, 12:21 PM
Hall Of Fame
F@H
 
Join Date: Dec 2010
Location: Fort St. John, BC
Posts: 1,222

My System Specs

Default

Yeah, I'm just curious because everyone uses such vastly different test benches. Some use the X58 platform and usually get around 3.8 Ghz, some use P55, and others use the P67... I also find that there is a huge difference in CPU speeds as there are clocks from 3.2 Ghz to 4.2 Ghz depending on the review.

Any thoughts on what CPU/clock speed is needed to remove a CPU bottle neck when you get into the high end SLI/CF realm?
Reply With Quote
  #4 (permalink)  
Old March 15, 2011, 12:26 PM
Hall Of Fame
 
Join Date: Feb 2011
Location: Ontario
Posts: 1,384

My System Specs

Default

Quote:
Originally Posted by draemn View Post
Yeah, I'm just curious because everyone uses such vastly different test benches. Some use the X58 platform and usually get around 3.8 Ghz, some use P55, and others use the P67... I also find that there is a huge difference in CPU speeds as there are clocks from 3.2 Ghz to 4.2 Ghz depending on the review.

Any thoughts on what CPU/clock speed is needed to remove a CPU bottle neck when you get into the high end SLI/CF realm?
It depends on the game but 4 cores at 4.2 ghz should be fine for almost everything.
Reply With Quote
Reply


Thread Tools
Display Modes

Similar Threads
Thread Thread Starter Forum Replies Last Post
best method to study? BrianHanna Off Topic 28 December 13, 2010 01:51 AM
What is your preferred method of Backing up your PC dustin1706 Storage 15 November 15, 2010 05:53 AM
Is there a lite method of testing new builds? YoungMan New Builds 5 November 13, 2010 09:13 AM
new scam method... CTA Off Topic 16 August 31, 2010 11:14 AM
New Mounting Method ?! Extreme_kid Air Cooling 11 October 6, 2009 11:04 AM