Asus' ROG Ares II: Four Dual-GPU Graphics Cards, Compared

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
[citation][nom]eric4277[/nom]Why did you include 7970 Crossfire and not 680 SLI?[/citation]

Like others have said before, two 680s in SLI are so close performance wise to a single 690, you can pretty much take those results as gospel.
 
[citation][nom]Onus[/nom]So sorry, but without FCAT measurements all of that work is for nought. I realize this article may have been researched and written before the significance of FCAT was realized, but that makes it an editorial task to not release it until it has been reworked, or even scrap it if it simply is no longer relevant.http://www.pcper.com/reviews/Graph [...] 70-Cross-6[/citation]

FCAT is just another tool. And until it's fully vetted as 1) not impacting results and 2) being accurate I would lean to heavily on the info it provides. Over the last two weeks everyone and their mother has hailed FCAT as a revolutionary tool, but I haven't seen any analysis into it's accuracy or it's effect on benchmarks. It wouldn't be the first time a tool gave inaccurate results. On top of that it requires much more complex testing. So I don't have any issues with it being left out for the moment.

On the other hand once it been cleared I would expect Tom's (and other reviewers) to adjust the benchmarks to to show full frames, partial frames, and time to render frames vs current benchmarks (there's nothing as misleading as an average).
 
So, just for clarity. You just posted a review of a card you can't buy anymore, and compared it to 3 other cards. One of them you can't buy anywhere either, the 2nd was never a real product and the third is the only one we can actually buy. Was this supposed to be your April Fool's post? But just like the news section here, a few days late.

This whole test is bogus. If you're going to compare multi-GPU cards that can't be bought anywhere, where the hell is the Bitchin'fast 3D 2000?

http://www.valken.org/BlogImages/bf.jpg

There is no way the Ares II can achieve 425 Bungholio marks. Sorry Asus, you only made it to #2.
 
Didn't know there are so many ati fanatics right here. Down voting everybody who mentions fcat or xfire failure? Seriously? I guess some people like stuttering.
 
Why publish a bar graph comparing information that doesn't start at 0??? Meaningless as a graph because graphically it conveys false information. Only by looking at the numbers and then processing the relative size of the numbers and purposely rejecting the actual "graph" can you understand the results.
 
So technically you could have 2 Ares II's in 4-way XFire? That would be, what, 1300W 100% load on just those cards? Yep, I want it.

Or are they not XFire compatible?
 
[citation][nom]Onus[/nom]So sorry, but without FCAT measurements all of that work is for nought. I realize this article may have been researched and written before the significance of FCAT was realized, but that makes it an editorial task to not release it until it has been reworked, or even scrap it if it simply is no longer relevant.http://www.pcper.com/reviews/Graph [...] 70-Cross-6[/citation]

Unfortunately, FCAT reviews take a lot of effort. I have a feeling we won't see a lot of them as a result.
 


AMD / ATI video cards have become credible contenders since the 7XXX series. They often match or even outperform their NVIDIA counterparts. Micro Stuttering when in SLI / Crossfire is inevitable no matter what GPU you decide to go with. It's less prevalent in current drivers, but the key word is less. It won't be totally escapable but they're at least making efforts to minimize it.
 
[citation][nom]g-unit1111[/nom]AMD / ATI video cards have become credible contenders since the 7XXX series. They often match or even outperform their NVIDIA counterparts. Micro Stuttering when in SLI / Crossfire is inevitable no matter what GPU you decide to go with. It's less prevalent in current drivers, but the key word is less. It won't be totally escapable but they're at least making efforts to minimize it.[/citation]
It was the 7000 series cards that they have tested with FCAT and they still have the issue unless you use a FPS limiter, v-sync or the CPU is a bottleneck.
 
Pcper won't attempt to run tests with vsync on or with radeon pro frame metering in place although both of which have been out for years / months as toms has highlighted in the past - rp makes a monumental difference to cfx frame smoothness. Until pcper actually do some testing with software that is available right now i find their articles irrelevant. My experience with dual 7970s is totally different to what they show due to vsync and rp DFC, so there reviews are totally irrelevant for me.

How hard is it to set vsync on , install rp and game ... not hard .. frame metering problem solved.

I guess pcper would have no news or make less headlines if they showed a couple of settings in existing software can fix 95 % of the issues.
 


Well, I for one don't see where FCAT or FRAPS latency could be inaccurate. Neither impacts benchmarking to any large degree and FCAT simply records the output sent to your monitor. I mean, taking just the output and measuring how many frames are in a single refresh wouldn't impact the results because those are the results. It wouldn't have any effect on the machine you're testing on either, because it's simply recording the video output into another machine. There might be under a millisecond of latency due to the DVI splitter but that's the best solution we have at the moment. FRAPS measures closer to the metal before the APIs and drivers take over, FCAT measures everything when it's gone through the entire rendering process.
 
On the first page of the article the author states "2 x 4.31 million" and "2 x 3.5 million" for transistor count. Don't they mean billion?
 
[citation][nom]Orthello77[/nom]Pcper won't attempt to run tests with vsync on or with radeon pro frame metering in place although both of which have been out for years / months as toms has highlighted in the past - rp makes a monumental difference to cfx frame smoothness. Until pcper actually do some testing with software that is available right now i find their articles irrelevant. My experience with dual 7970s is totally different to what they show due to vsync and rp DFC, so there reviews are totally irrelevant for me.How hard is it to set vsync on , install rp and game ... not hard .. frame metering problem solved.I guess pcper would have no news or make less headlines if they showed a couple of settings in existing software can fix 95 % of the issues.[/citation]
Pcper is the only ones who have done so with FCAT:
http://www.pcper.com/reviews/Graphics-Cards/Frame-Rating-Dissected-Full-Details-Capture-based-Graphics-Performance-Tes-11

And Radeon Pro does not add frame metering, or at least the only tests done on it have shown it to be a form of v-sync and FPS limiter, so it has only been tested in that context.
http://www.tomshardware.com/reviews/radeon-hd-7990-devil13-7970-x2,3329-11.html

EDIT: I notice you referred to DFC, or dynamic frame rate control, which by RadeonPro's definition, it is a FPS limiter. So yes, when FPS are limited, Crossfire works pretty well. Pcper has also made not that when FPS are limited due to CPU bottlenecks, it seems to force some spacing as well.

So if you are good with FPS limiting, Crossfire is fine.
 
[citation][nom]damianrobertjones[/nom]"the GeForce GTX 690 remains the most elegant high-end card we've ever used."I actually had a chance to own a 690 so I went for it and, upon the next reboot, I was greeted by a black screen. Yay. I looked on the web and, as expected, others were suffering with the same issue. An ebay auction later and I'm back to 2x 660s in sli. The moral of the story is: Just because it's expensive doesn't mean it'll work[/citation]
lol, so you have one bad card and you will disregard that card alltogether? a very small sample size compared to all the ones that do work. Just fyi there are people with 660's not working too. Are you now going to sell your 660's?
 
Still would have been NICE to bench the card with the NEW drivers as well as the old ones that way they could still compare to the cards that they had to send back to the manufacturers. Don't know why I got down voted for saying that above, it makes a lot of sense to use the new drivers as well as the old. They used old drivers for the 690 as well... So basically the whole review used OLD drivers for direct comparisons but it does not represent current performance for any of the cards..
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom