Welcome to Dual GPU world. If you want FPS get a better single GPU card and you are almost 100% free problem mate
The only games that have problems with stuttering that I "play" are rFactor 1 and 2, I guess the problem isn't the SLI. To get the performance for triple screen(in the near future)I needed a GTX770 that costs a lot more than adding a 2nd gpu.
If adaptive vsync is actively running you will have lag just like with normal vsync. Some either don't notice it (the ones that state they have no lag) or aren't bothered by it. If you ever play without vsync and can avg around 100+ fps you won't go back to having it set. Only way to experience AC (AC runs great at 120fps) and rF2 IMHO. There has been a number of threads discussing running higher FPS than your display can do. Doesn't sound intuitive but it feels like a different sim. Best advice if you want to enjoy rF2 the way it should be is sell both cards and get one. I've been kicking around this forum for awhile and one solid GPU seems to be the way to go. Would be nice if that changed but ISI has been very quiet about any plans for linked GPU improvements.
If you're having problems in rF1 as well then I wonder if you've got another issue. I run two GTX 670 SC in SLI. rF2 performance is acceptable (60-120 fps) on three screens with Multiview on and HDR off. In rF1 though depending on the mod and track I get framerates between 120 and 400+. I had to force a framerate cap at 150 because the super-high framerates were crashing my card. I suppose there's a little micro stuttering from time to time but nothing that concerns me. In rF1 ensure your shortcut has the +fullproc switch... that makes a huge difference for me.
Hi, I have been busy in these past days but I had the opportunity to test a few things: I tried different clocks and same clocks in the 2 gpu's. I also tried different configs in both configurations. What I have noticed is that with the same clocks in the cards the stuttering is a lot less compared to different clocks, but the stuttering is still there. A very strange thing is that in Silverstone when the Loading ends it takes a lot to advance to the next phase, and when I'm able to click in RACE the mouse and the computer starts to kinda block and get unresponsive, very strange. I don't have this problem in any other track, for example in Asseto Corsa I have 0 problems and good fps, no stuttering at all, so it's not a hardware problem.
You are most likely right that there is still going to be some lag with adaptive vsync on. However, if it is so small that it's neither perceivable to the human eye (i.e. in the very low ms) and does not negatively affect your performance the slightest bit but instead helps with stutter and screen tearing (as is the case with me and all others who use it except the OP) then it makes zero sense to suggest you shouldn't use it. Now the OP says when using it he can clearly see the input lag and there's something really wrong with this observation that tells us there is a problem on his system because you simply should not be able to see and/or feel any input lag whatsoever.
Maybe I'm misunderstanding you but because you can't perceive input lag, and have the incorrect assumption that "you simply should not be able to see or feel any input lag", anyone that does has a PC issue end of story? I can clearly see input lag with vsync on and there is nothing wrong with my PC, OC 4770K, etc. I've seen it on other PCs as well. There has been plenty (plenty) of threads discussing lag with vsync, would you suggest all these folks have PC issues? Most can tell there is lag very easily with a typical 60fps display and vsync on, they put up with it because depending on your GPU (as you said) the sim might not run smoothly without. If you have a refresh rate on your screen of 120fps and are using vsync then that is another story. If not and you are at 60fps do you mind sharing how you test for lag?
This wasn't just my own experience but that of many others who are also using adaptive vsync. However, you raised a good point regarding 60fpz/hz monitors, something that i forgot to consider because at 120hz with adaptive vsync it's pretty much perfect and not perceivable to the naked eye. So i ran a test of my own with my monitor at 60hz... The video is also recorded 30fps (so every frame lasts for 2 whole frames in-game) so any perceived lag is appears more severe than it would in real life. To my eyes, i would say that a very subtle amount of extra input lag is noticeable more with adaptive vsync than no vsync but only if your looking for it by performing such tests. Is this the same amount of perceived input lag as what you experience with adaptive vsync at 60hz? I also performed the comparison at my monitors native 120hz but the results are so identical that i didn't want to waste time uploading them (also because my upload speed is super slow..only 100kb/s). I found an iracing with adaptive vsync video (skip to 3:26) and it looks a lot worse on it for some reason.
It wasn't my intention for you to spend a lot of time on this but to recap we started at: "...you simply should not be able to see and/or feel any input lag whatsoever" And now: "To my eyes, i would say that a very subtle amount of extra input lag is noticeable more with adaptive vsync than no vsync..." We are almost on the same page, however the comment about only seeing this in performance tests doesn't account for players that use the cockpit view with the wheel moving. I go back and forth depending on the mod. To answer your question, that looks similar to what I see when I use vsync. I have an old video on my ipod and it looks a tad worse but not by much. I play on a much larger screen (65" projected image) so I have almost the entire wheel in view. When racing and making quick corrections throughout the lap the input lag is clearly noticeable. What is odd about your vid is you still seem to have lag with vsync off but I didn't see the FPS shown when it was disabled. My point to the guy who started the thread is the same, after living with input lag I came across some threads about running much higher fps than the display could handle with vsync off. I gave it a try even though it didn't make a lot of sense. I couldn't believe the difference. Not only was there no lag the sim itself felt a good leap better. Unfortunately, running with one solid GPU is the best option for this. I try to have things set to where I don't drop below 80fps, avg 100+ and I cap the max in the plr at 120. Each to their own but in my experience (and others) it makes rF2 that much more enjoyable.
I need to apologies about the previous video. After you raised a good point that there is lag in both comparisons i went back to test again without adaptive vsync. It turns out i also had it enabled in the programme specific settings for rf2 in the nvidia control panel, so disabling it at global settings did nothing and both comparisons actually had adaptive vsync enabled. Really sorry about that, lol. So i tested with adaptive vsync off and the first thing i noticed was that in-game fraps was no longer reporting a solid 60fps but 180+fps. That should have been my first clue that my setup was wrong but i thought it might have just been artificially capped in the plr file. My (very dumb) mistake. Well, your absolutely right in saying there is input lag with adaptive vsync and it's made extremely obvious compared to when it's disabled at 60hz. So i apologies for my brash/flippant comment that it should not be able to see any input lag at all. I retested at 120hz and there is an ever so slight amount of input lag but it's no where near as self-evidence when running at 60hz. You can notice it from time to time in this video at moments when the wheel is reversing direction extremely quickly. So even at 120hz it's there compared to when adaptive vsync is disabled, albeit many times more subtle compared to when running at 60hz...but it's still there. Sorry to have wasted your time.
You say your performance is acceptable but have you ever run your system with just one card for comparison? A lot of people (most people?) seem to get worse, or at best little to no FPS improvement with a SLI configuration. Count me in that group. Not sure why NunoPinto has trouble with rF1 but problems with SLI in rF2 are unquestionable. And my forecast, after two years of the community pleading to no avail for this to be made right, is that it is not going to get better... for whatever possible reasons, which are too numerous to list. Those who can fix this just don't seem to care.
No apologies needed and not a waste of time at all. Hats off to you for approaching this in a very logical way. Cheers.
Guys you know that input lag heavily depends on what monitor you use? So the experience will differ (much). Greets Pete
Many possible solutions and or workarounds are now posted in this thread. As you can see the point is you have to made a compromise with SLI and deal with the disadvantages less stuttering but more input lag or vice versa. @ the moment my SLI solutions is OK for me but if I look exactly to my rF2 I'm not free of micro stuttering but that is the compromise I had to make. If nothing changes my next card will be a strong single gpu card. I've said this once it's sad that I must spend so much money to a hardware company but if ISI and or nvidia can not deliver a solution I've no choice. (I don't want do without rF2 )
Here you can find many additional informations I know it is a lot of reading but for a real sim racer this is a gold mine. http://isiforums.net/f/showthread.php?t=17996
Refresh rate its not the main reason of input lag with or without adaptive vsync. The monitor itself is the key. I'm pretty sure that you can have a 60hz monitor with less input lag than other with 120hz. I have 3 22'' monitor to play, 2 of them old LG and the 3rd are HP from professional line that i bought because it have displayport because it was the only solution to connect my old 5870 with the other two and even that way the HP being much more expensive and having an enormous diference in colors than the other two it have much more input lag and i need to put it on side instead of being in front of me. All depends of quality of the monitor.
I have a strange problem with running 2x780 sli. Every track EXCEPT for Portugal run great even Silverstone with Honda. At Portugal I get about a one second pause at certain parts on the track, no matter what car I drive. If I disable sli there is no pause and everything runs great. Any ideas?
Have you tried using something other than display port to connect that HP monitor? Can you do a test with just that monitor hooked up as a single via DVI-D with vsync at 60hz then do the same test with your LG? Would be interesting for the community to see the huge difference. I've used a fair amount of different displays at 60hz and (with vsync) all had about the same amount of input lag. Even my 60hz LED projector matches the input lag pretty well with DrR1pper's monitor at 60hz with vsync enabled. I'm not disagreeing that monitor type makes a difference but to state, "I'm pretty sure that you can have a 60hz monitor with less input lag than other with 120hz" is complete nonsense if you are using vsync on both at their native res. DrR1pper made some great vids to show the difference at 60 vs 120. That is consistent with everything I've seen. If you can prove your statement make some vids of your own but use dual link dvi for your 120hz tests. Actually to be fair you should use that on all monitors you test with for consistency.
I am running the latest nVidia drivers 332.21 on a GeForce GT 650M SLI config, with the SLIs overclocked (GPU clocked to 1050, memory clocked to 2250). I run 4x AA, vsync on, max framerate 62 fps in my player.plr, and I get no stutter with SLI, and a good capped framerate of 60fps. I do get a slight wheel lag, as has been noted by everyone who runs with vsync on, but I rather a smooth screen than tearing that I no vsync and even adaptive vsync still produces...
