Crossfire or upgrade?

[Disposed]

I have an xfx 7950 just finished my eyefinity setup and its become clear I need more horsepower. Question is should I grab another or buy a faster card and sell mine while it still holds value? 1080p monitors, more than enough power for a dual card system just trying to make my mind up. Any input?

 

[Flopper]

crossfire, any other card is a Titan.
or lower settings and OC current card.

 

[Disposed]

crossfire, any other card is a Titan.
or lower settings and OC current card.

Titan is shit, 2 7970s are faster for a lot less money. Unfortunately my card don't overclock for shit.

 

[harmattan]

If you haven't read the article on AMD's issue with frame rendering in Crossfire you should. HardOCP knew this from observation real world testing but couldn't put their fingure on the issue, I knew it from my working with many different SLI and Crossfire configs (some side-by-side), and now the industry has developed a testing toolset that demonstrates it. In summary, Crossfire does not render entire frames in some instances (although "normal" tests/benches will count this frame in results) which, in a word, makes the experience quite shit.

Go with a single card or SLI.

 

[Disposed]

If you haven't read the article on AMD's issue with frame rendering in Crossfire you should. HardOCP knew this from observation real world testing but couldn't put their fingure on the issue, I knew it from my working with many different SLI and Crossfire configs (some side-by-side), and now the industry has developed a testing toolset that demonstrates it. In summary, Crossfire does not render entire frames in some instances (although "normal" tests/benches will count this frame in results) which, in a word, makes the experience quite shit.

Go with a single card or SLI.

that's what I have heard. how is Nvidia triple monitor compared to eyefinity?

 

[EPOQ]

If you haven't read the article on AMD's issue with frame rendering in Crossfire you should. HardOCP knew this from observation real world testing but couldn't put their fingure on the issue, I knew it from my working with many different SLI and Crossfire configs (some side-by-side), and now the industry has developed a testing toolset that demonstrates it. In summary, Crossfire does not render entire frames in some instances (although "normal" tests/benches will count this frame in results) which, in a word, makes the experience quite shit.

Go with a single card or SLI.

It's funny that one site does a LIMITED test and a bunch of people jump on this bandwagon that AMD multi-card setups are completely worthless.

Let them do some more in depth testing before making the assumption that Crossfire is "quite shit." I'm pretty sure if this were the case, you'd see hundreds or thousands of CF users complaining like crazy about how they can't tell a difference between having 1 or 2 cards installed. I'm seeing the opposite from many users on this forum, as well as others. They are employing the use of frame limiters and other tweaks that PCPer wasn't using, so I would say wait and hold off to see what further testing shows. Also, if this is brought to the attention of AMD (I'm sure they've already seen it) -- we can expect some driver fixes in the near future.

 

[Disposed]

I'm not worried about that if there is really an issue I would expect a driver fix eventually. Now that I have read the article I'm not totally convinced anyway. Who knows what their device is really showing. If Kyle and the gang look into it and find the same I would take it more seriously.

Thing is I'm just not sure if it's worth buying another 7950 when I could buy a couple 7970 or 670s and sell the 7950. I'm really digging eyefinity and don't want to look back and wish I just bought better cards while my 7950 still had value.

 

[Disposed]

So I just said fuck it and bought a pair of evga superclock 680s. If my 7950 could hold a stable overclock I would have lived it instead it would tease me and play at 1100+ but would eventually crash so I could see the massive increase in performance but not use it. Overclocked it's fucking amazing stock its meh.

 

[PhantomTaco]

It's funny that one site does a LIMITED test and a bunch of people jump on this bandwagon that AMD multi-card setups are completely worthless.

Let them do some more in depth testing before making the assumption that Crossfire is "quite shit." I'm pretty sure if this were the case, you'd see hundreds or thousands of CF users complaining like crazy about how they can't tell a difference between having 1 or 2 cards installed. I'm seeing the opposite from many users on this forum, as well as others. They are employing the use of frame limiters and other tweaks that PCPer wasn't using, so I would say wait and hold off to see what further testing shows. Also, if this is brought to the attention of AMD (I'm sure they've already seen it) -- we can expect some driver fixes in the near future.

FWIW keep in mind the placebo effect or expectations theory, when you expect it to do amazing you can convince yourself sometimes that it is amazing. It does sound like I might be saying people are stupid and not noticing, but that's not my intention. It's a documented scientific phenomenon, and it may very well apply here. Also I've never thought in my time using both companies that AMD ever seemed smoother than NVIDIA, but always the opposite. I am a fan of NVIDIA though, and do acknowledge there may be a bias. Also applying all these tweaks and stuff is indicative that there is a problem at stock don't you think? And the fact that it took until review sites pointed it out for AMD to address it doesn't speak highly of them either. Regardless, I'm glad it's come to light and I hope AMD fixes it (if it is in fact an issue) soon, in the end I want both companies to be successful so I have as much choice as possible when buying.

 

[Fooshnik]

In summary, Crossfire does not render entire frames in some instances

This is complete bs. Two frames will hit the display driver at nearly the same time, causing the first to be displayed truncated by the second with vsync off. Most people using crossfire use a frame limiter to mitigate microstutter at which point this issue is moot. With vsync off, every frame is truncated to some extent when frame rate exceeds refresh rate.

It's sad the authors of the article didn't clarify what they're showing.