Frame Rating and AMD crossfire issue

[Roooker]

Some of you might know the site www.pcper.com. In case you do not know it check it out. They so great Hardware rewievs and stuff.
So back to the topic:

AMD Crossfire is a lying at frame rates.

Simple as that. Dont believe me ? Check this rewiev. Please read it seriously because it is quite complicated.

Here

Just wanted to know what you guys think and maybe we can get Tom to make a vid about that, because i think it is a pretty big issue.

Here is just an idea on what the whole issue is about:

Everyone until now has seen that as the results. We see that a 7970 GHz Edition is little better that the 680. Also if we do the crossfire and sli comparisons.
BUT now look at this graph:

That the the result of the exact same test just somehow different. Take a look at the 7970 crossfire result. Almost the same as a single 7970.
So what observed means is ... this is taking too long to explain just watch the video Everything is explained there.

 

[grassman]

If anyone didn't notice then surely they are brain dead. I way playing Far Cry 3 with 2 7870's and the thing was just a cluster #&%@ of micro stutters Fraps would present 30 FPS but when you are actually looking at it it's actually half of that :\

Dual cards suck arse period there is a reason why gen after gen people just buy a new single card.

 

[s3ns3]

well.this sucks

 

[alxcsb]

So, we have to trust a guy, with a home made overlay, some dodgy second system that records video from, supposedly, the same source, and after all that, we have to trust that all that info and data is processed without any bias by an nVidia made tool?

 

[nepas]

So, we have to trust a guy, with a home made overlay, some dodgy second system that records video from, supposedly, the same source, and after all that, we have to trust that all that info and data is processed without any bias by an nVidia made tool?

errrr,no,seeing as it is showing everywhere you look these days

 

[s3ns3]

its not the first guy/site i see with this new ''methodology'' . all saying the same

 

[grassman]

None of them mention that it's a stupid investment on either side.

 

[Deceit]

I am using two 7970's and up until the 13 series drivers, micro stutter was horrible in crossfire.
But now the issue of micro stutter was pointed out to Amd they've seemed to have got there act together and my gaming experience is buttery smooth on 3 screens at 5760*1200 .
I was tempted to switch to Nvidia because of this but now there is no reason too and I have found no erratic drops in frame rates.

 

[Roooker]

@Deceit It is not about frame drops. It is about driver sending out a frame that is 5 pixels small. Fraps counts that as a frame but you dont see it as a real frame on the monitor, because it is only a few pixels high. So fraps is reporting a "faked" number of fps.

 

[alienware]

I am using two 7970's and up until the 13 series drivers, micro stutter was horrible in crossfire.
But now the issue of micro stutter was pointed out to Amd they've seemed to have got there act together and my gaming experience is buttery smooth on 3 screens at 5760*1200 .
I was tempted to switch to Nvidia because of this but now there is no reason too and I have found no erratic drops in frame rates.

If I had a pound for every time some one said AMD have got their act together with drivers I would be a rich man.

If anyone didn't notice then surely they are brain dead. I way playing Far Cry 3 with 2 7870's and the thing was just a cluster #&%@ of micro stutters Fraps would present 30 FPS but when you are actually looking at it it's actually half of that :\

Dual cards suck arse period there is a reason why gen after gen people just buy a new single card.

I don't agree with that. I will make some videos of me playing all of my games perfectly smoothly on my rig. I've had Crossfire a couple of times and both times it was absolutely crap. Every time I've ran SLI I've had minimal issues that are taken care of pronto.

Edit in fact I don't need to. Watch the video.

 

[MrKambo]

i watched this video when it first came to the web, and tbh from my experience, i agree that SLI seems better than Crossfire, but when it come to cost vs performance crossfire makes more sense

from what ive read, the 13.3 drivers are starting to fix alot of these frame rendering issues, i havent had a chance to try them out yet, but because someone took the time to make this problem visually im sure we will see fixes from AMD

 

[alienware]

The latency sucks on Crossfire aswell. It's that causing the micro stutter.

The only SLI set up I've had micro stutter on was Quad SLI with two 295s and that was because it wasn't supported properly and trying to run in normal SLI.

I did find a fix for it though, not that I would ever run it again (got the cards free sponsorship).

http://www.tomshardware.com/reviews/graphics-card-benchmarking-frame-rate,3466.html

Tom's are on it too....

 

[grassman]

Its k guys AMD are buying up the software market such as other noble company's like Intel and Nvidia.

Your issues will be resolved soon..

 

[alienware]

Its k guys AMD are buying up the software market such as other noble company's like Intel and Nvidia.

Your issues will be resolved soon..

:lol: hahaha.

I really hope that guy Ryan pulls out a set of 5xxx series and shows what was happening on those..

I had 5770CF and it SUCKED.

 

[MrKambo]

id love to hear Tom's take on this...

 

[iBeInspire]

There goes every hope I ever had of Crossfiring my HD6850.
I'll save up and get a mid range 8000 series, all I'm hearing these days about crossfire are ''problems, problems, problems''.

 

[alienware]

id love to hear Tom's take on this...

Their article is pretty infantile at this time. Give them time and no doubt they'll start knocking out results.

A few websites have the required hardware now.

At least I don't feel like a dickish fanboy for saying that Crossfire sucks and SLI kicks its ass any more.

Got it all in black and white now !

 

[Deceit]

If I had a pound for every time some one said AMD have got their act together with drivers I would be a rich man.

I've had 30+ graphics cards in the last 5 years (mix of both nvidia and ati) and other than this crossfire issue I havent had a single issue with either nvidia or ati drivers. Maybe I'm lucky but I wouldnt complain about the quality of Amd's drivers from experience

 

[alienware]

I've had 30+ graphics cards in the last 5 years (mix of both nvidia and ati) and other than this crossfire issue I havent had a single issue with either nvidia or ati drivers. Maybe I'm lucky but I wouldnt complain about the quality of Amd's drivers from experience

There are a lot of issues to be had with AMD drivers. This is where the "But the latest drivers part the red sea" sort of stuff comes from.

I don't know exactly who is taking care of the drivers now tbh. Could be the original ATI team but AMD could have well replaced them.

I've always had a passionate hatred of Crossfire. Mostly because I believed all of this "But the latest drivers" bull poo but at the end of the day it was simply rubbish.

I really hope they go back and test older cards and older drivers tbh. I'd love to see if I was duped. However, it doesn't matter because real world experience says it all. I hated every minute of Crossfire yet I've had numerous SLI systems and never once developed the hatred for SLI.

I thought it might be because Nvidia drivers have given me hardly any crap but now I know the real reason.

At that time it was common 'knowledge' that Crossfire scaled better than SLI. IE - by adding another card you got a much higher % increase in frames. What I would love to know now is if they were runts and dropped frames. That is, tbh, the only real explanation as to why it was so buggy.

 

[Roooker]

I dont understand why there is such a big discussion here. If you watch the 20 minute long video, Ryan (the guy in the vid) explains everything. Me mentions how they measure everything and how the analyse the data. He even explains the results they got, like frame time veriances and stuff. If you have understood once what this method is about and what the difference is to the normal fraps fps counter the only importent thing are the graphs. And pcper.com got lots of them showing serveral games on multiple resolutins with a whole range of cards. And i every game on every resolution with every gpu the crossfire result was the same.

Let me quote Ryan here:
"AMD is aware of this issue, we have talked with them serveal times about it and they promied us that a solution is comming although propably not for a 60, 90 to 120 day to actually get a global kind of fix for this type of thing."