[OCClub] Crossfire vs. SLI Performance - Page 3 - Overclock.net

shadowk · 04-28-09

Quote:

Originally Posted by Robilar

Shadowk, nice to see that you put some effort into your response.

Here is the thing:

I read all of those reviews when they first came out. In fact two of them influenced me to buy my pair of GTX295's (as there was only one other member at OCN at the time with one).

Upon getting the cards, I spent several weeks benching, a couple of reformats, trying multiple driver revisions in order to get the same kind of results the web reviews posted.

I also spoke to members here with the card (and mates of mine at XS) in order to see if there was some issue.

I'm at the point where I take every review site with a grain of salt. How exactly do you think they get their hardware? manufacturer's supply them with comped equipment for testing purposes.

It would be foolish for any site to post all glowing reviews (although hardware canucks comes pretty close) as it would discredit them but there are many ways to spin negative aspects of hardware into a positive result.

Imagine if a site were to post truly negative reviews on hardware that deserved it? sites like HardOCP do that from time to time (their foxconn bloodrage review was brutal...) but I expect that even they have to be careful of crossing the line to far.

My results were inconsistent with most reviewed sites. When I asked other members here for their own test results, there was parity to what I was seeing as well.

Far be it for me to believe that manufacturer's cherry pick parts they send out for review (Intel for example would never send out an ES processor that overclocks much higher than production retail release product....) but I have to wonder why real world tests (Read site's like OCN that have no affiliation to hardware manufacturer's) differ so much from sites that actually review hardware as part of running a business (namely their website)?

make sence

so basicly i cant trust any hardware review (that a sad world we live in ) (with the exaption of INTEL )
thank you for taking the time to explain you'r point +1

OpTicaL · 04-28-09

Quote:

Originally Posted by Robilar

I don't think I have ever seen such an nvidia slanted review before...

All reviews are financially motivated. If AMD was backing OCClub then the review would've been slanted towards ATI's side. Sad thing is, many people take these reviews SERIOUSLY and draw themselves into a endless spiral of mindless raging.

Reviews by Robilar are neutral but we don't have many members here that have the funding and neutal stance as Robilar has on OCN.

Quote:

Originally Posted by shadowk

fine then im crazy

Prime example of someone taking review sites, SERIOUSLY.

RussianGrimmReaper · 04-28-09

Quote:

Originally Posted by Robilar

You are kidding right?

I listed the games I tested above

COD4
COD WAW
Grid
L4D

Last time I checked none of those games was optimized for ATI or nvidia

Furthermore, as I mentioned above there are numerous reviews that put ATI ahead of Nvidia in some apps at some resolutions. Its basically a wash in many cases.

Grid is a very ATI optimized game. It always has been. I'm an ATI fanboy, I would know

CowColor · 04-28-09

nVIDIA FOR THE WIN.

bumsoil · 04-28-09

gtx 285 SLI~=4890 CF > GTX 260 SLI > gtx 295 > 4870x2

GSingh · 04-28-09

Quote:

Originally Posted by CowColor

nVIDIA FOR THE WIN.

lol. They got him too.

angrysasquatch · 04-28-09

http://www.overclockersclub.com/revi...e_vs_sli/6.htm

Lol, the 4870x2 cf loses 1 fps from going up to 2560 from 1920. And in the 1920, the 4850x2 cf pulls ahead of the 4870 x2 cf? It's pretty ovbious the test needs to be run again.

003 · 04-28-09

Quote:

Originally Posted by Robilar

For example with COD WAW using the i7 board, cpu at 4ghz, ram at 1600mhz, video cards at stock:

Pair of 4890's averages 225 FPS
4870X2 averages 175 FPS
Pair of 260's averages 155 FPS
GTX295 averages 130 FPS

at 1920x1200

Rob, I saw an article a while back, I believe it was on Tom's Hardware, although I could be wrong. Basically, it found that there seems to be some kind of performance problem with nvidia GT200 based cards on the i7 platform, and that they will get outperformed by ATI/AMD alternatives which would previously have been much faster on Core 2.

I have been trying to locate this for a while now and it's driving me nuts because I can't find it. Although, I believe it holds a lot of credibility because I do not believe that my GTX280 SLI setup is performing to it's best. There is some flaw in nvidia programming on i7.

I want to see a good multi GPU comparison on a high clocked yorkfield system, it would be really interesting...

dr_bowtie · 04-28-09

Quote:

Originally Posted by -iceblade^

Fail.

don't even need to read the source - if he is claiming this then he has no knowledge of wha is going on. ATi has been leveraging Crossfire as a means of getting cheap bang for buck performance that often takes out nVidia's top parts for a fraction of the price since the 3850.

http://www.anandtech.com/video/showdoc.aspx?i=3151&p=10

the age of the $1000 video card is over, and multi GPUs is a viable solution for anything 1920x1200+: you don't need to be splashing out on a 2560x1600 monitor to be able to benefit from it
what a bad article

I dont get where you are comming from here...they are showing you all of those card in Xfire and SLI...from a single 4850 to a 4850X2 in xfire as well...

Basically what he states in his tid bit is you need a 24" up monitor to even realize the horsepower potential of the cards...

even with a 22" monitor a 4850X2 is overkill...it will pull more FPS than your monitor can show... most monitors have a refresh rate of 60hz and anything more by the way of FPS goes unnoticed...

basically why shell out big money for a video card if the monitor you have cant realize the potential of the card....just means the money you spent for the power you want never gets used...

Quote:

Originally Posted by angrysasquatch

http://www.overclockersclub.com/revi...e_vs_sli/6.htm

Lol, the 4870x2 cf loses 1 fps from going up to 2560 from 1920. And in the 1920, the 4850x2 cf pulls ahead of the 4870 x2 cf? It's pretty ovbious the test needs to be run again.

actually that test was run 5 times by both parties and they got the same results...it's just one of those fluke things...

Quote:

Originally Posted by OpTicaL

All reviews are financially motivated. If AMD was backing OCClub then the review would've been slanted towards ATI's side. Sad thing is, many people take these reviews SERIOUSLY and draw themselves into a endless spiral of mindless raging.

Reviews by Robilar are neutral but we don't have many members here that have the funding and neutal stance as Robilar has on OCN.

Prime example of someone taking review sites, SERIOUSLY.

this couldnt be farther from the truth and is completely wrong...if you look back on the site last year at this time they were praising ATI when the 4870 was the top dog....this site isnt financially motivated by ANY supplier...their parts are funded by the site owner Neoseeker....

RussianGrimmReaper · 04-28-09

Quote:

Originally Posted by bumsoil

gtx 285 SLI~=4890 CF > GTX 260 SLI > gtx 295 > 4870x2

No... Just no.

Quote:

Originally Posted by angrysasquatch

http://www.overclockersclub.com/revi...e_vs_sli/6.htm

Lol, the 4870x2 cf loses 1 fps from going up to 2560 from 1920. And in the 1920, the 4850x2 cf pulls ahead of the 4870 x2 cf? It's pretty ovbious the test needs to be run again.

GDDR5 has massive latencies. GDDR3 does not. Unless a game is memory bandwidth critical, GDDR3 will outperform GDDR5.

		Overclock.net - Overclocking.net > Industry News > Hardware News
[OCClub] Crossfire vs. SLI Performance

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)