![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
AMD should be held responsible to CF owners who bought cards based on reviews.
- Thread starter tajoh111
- Start date
- Status
Tamlin_WSGF
2[H]4U
- Joined
- Aug 1, 2006
- Messages
- 3,074
It is the hardware frame metering I'm speaking of. I did think 680 had it too though, so my mistake there.
There has been complaints about microstutter on every generation of SLI before Keplar, which is my point. According to OP, Nvidia should then also be held responsible to any owner who purchased SLI and experienced microstutter.
OP is creating a big stink on this, since Nvidia has improved its multi-GPU configurations regarding microstutter. But, he is neglecting to mention or is ignorant about that this has been an issue on Nvidia cards also for years.
Then how do you explain 680 SLi results which lack hardware metering? You are right though, afaik hardware metering was first introduced with kepler on the 690.
Frame metering has been around since G80 though, if not earlier, and I think you will find results that show SLi had less ms even back then. Yes it has progressed since then too, though. I've owned 3 CFX systems (3870, 4870, 5850) and I do remember 8800GTX SLI being smoother, and I had to use the same work arounds people are talking about now for 7970 CFX (frame limiters, vsync etc).
Frame metering has been around since G80 though, if not earlier, and I think you will find results that show SLi had less ms even back then. Yes it has progressed since then too, though. I've owned 3 CFX systems (3870, 4870, 5850) and I do remember 8800GTX SLI being smoother, and I had to use the same work arounds people are talking about now for 7970 CFX (frame limiters, vsync etc).
I agree, according to the OP people who have been enjoying crossfire had somehow not been enjoying it the entire time.
Microstutter has been present before but has typically been worse on AMD hardware. People have been aware of microstutter for a while and people buying dual cards accept that.
Microstutter isn't the big issue here. People have accepted microstutter is present with a dual card setup.
However what I am talking about CF scaling being none existent, to scaling 20 to 40 percent, in games, when a FPS counter like FRAPS(the tool of reviewers) indicate otherwise. Basically 1 out of 2 frames are being rendered on the screen. 1 full frame and 1 runt frame which might be 5 lines on a screen. However a program like fraps counts this as two frames and thus tricks the FPS counter. Your missing out on 1 out of every two frames, this is a big deal.
That basically means your getting a CF experience with no issues in only 1 out of 6 games when combined with other issues. People couldn't detect this with tools before. Now they can. What they could detect is that they were not getting a smooth enough experience when FPS were pushed low enough.
Some people likely not noticing this issue, may be gaming on a single monitor and are getting pretty close to a playable experience off a single card and adding another card adds that extra 20 percent, that makes it a tad smoother. Other cases, vsync seems to be helping(but with the cons associated with vysnc). However when people read reviews, they are looking for fastest setup(as they don't test with vsync on) and this is where AMD has lead them astray.
Why get a crossfire setup if your getting issues in 83 percent of cases. Your betting off sticking with a single card or getting a Nvidia setup.
Getting framerates that are actually half or 70 percent the figure in what you would see in a a review in half of the games(AAA titles at that), should be a big deterrent from people getting a CF setup . If people knew they were not going to get crossfire scaling, would they have still purchased a dual card setup. I don't think they would have.
This issue should be seen as much as possible and having two threads that highlight this weakness are doing a service for gamers.
These being potential owners should think twice about getting a crossfire setup because of these issues and are for the most part wasting their money if they decide to get a second card. And current owners should be compensated because they were already were kind enough to AMD for forgiving the misgivings of CF that were initially published(the problem where it didn't scale in 1 in three games). Now with new reviews showing it doesn't scale in 1 out of 2 games(possibly independent from the first set up of games), leaving potentially only 1 in 6 games working, these owners really got screwed over.
People who purchased an ARES II card particularly got screwed over. They paid 1500 dollars for a card, which supposedly had the moniker fastest card in the world. With these issues present, that is not the case and they paid 1500 dollars for a setup with a sub par dual card experience.
This will be my last post here in this thread, because I already provided evidence over and over. People just seem to ignore it and say CF rocks, these problems are insignificant or OP is a shill. This is simply irrational when this issue has this much evidence and AMD has already admitted to it.
YeuEmMaiMai
Fully [H]
- Joined
- Jun 11, 2004
- Messages
- 32,569
tajoh111, quit being ignorant, seriously
AMD drops a crapload of frames into the buffer in a short time. the each successive frame gets displayed resulting in tearing.... hello? turning on vsync fixes the issue.. because it only allows a frame to be presented during the blanking signal
AMD drops a crapload of frames into the buffer in a short time. the each successive frame gets displayed resulting in tearing.... hello? turning on vsync fixes the issue.. because it only allows a frame to be presented during the blanking signal
- Status