Would you exchange 2x7970's for 2x GTX 680's?

[icemanchilled]

Seriously, iceman, you're saying that while inputlag from other sources is known problem, but suddenly when it comes from nVidia solution, it suddenly doesn't account into it at all?

No, that's not what i'm saying at all. Firstly, input-lag is always going to be an issue, the physical nature of the hardware virtually ensures that. But secondly, the caveat to this is that input-lag only becomes an issue when it becomes significantly high. So, hence my repeated question: Where is the evidence that the input-lag introduced by this fix is sufficiently high enough to affect your overall gameplay experience? We have solid evidence of micro-stutter affecting gameplay, but we have zero evidence that the input-lag is sufficiently high enough to cause problems. Clear enough?

I only indicated there's people who have had / have problem with input lag, and question how many of those might have been partly because of this tech, not that it definately is. If you change your parts due inputlag, of course every single thing accounting into it matters, regardless if you know the source or not.

Again, see my paragraph above. I accept what you're saying, but where is your evidence that the input-lag significantly contributes to this CB? You make these claims without actually having any evidence to back it up, you just keep saying it must be bad, because input-lag is bad, MmmmKay?

[drinksalot]

Can't be bothered to read the whole thread so it might have already been mentioned, but doesn't Kepler now support hardware based frame rate metering?

Everything I've read say the SLI'ed GTX 680's offer an extremely smooth and responsive gameplay experience.

[codydewey]

Can't be bothered to read the whole thread so it might have already been mentioned, but doesn't Kepler now support hardware based frame rate metering?

Everything I've read say the SLI'ed GTX 680's offer an extremely smooth and responsive gameplay experience.

Quoted for truth. Not sure about the monitoring, but smooth and responsive would be accurate.


Not sure why we have all this speculative crap flowing into this thread. Wait, just realized the person who was posting it.

Welcome back CB.

[Kaotika]

Can't be bothered to read the whole thread so it might have already been mentioned, but doesn't Kepler now support hardware based frame rate metering?

Everything I've read say the SLI'ed GTX 680's offer an extremely smooth and responsive gameplay experience.

They've had the hardware there at least since G80, microstuttering wasn't just big enough issue talked about before 5xx-series, when they first talked about it. Now as it's talked about around, it was convenient time to actually give name for it, too.

SLI'd GTX 680's give smoother than CFX solutions, but less responsive as input lag grows due their solution reducing input lag. (note: less responsive doesn't mean it's "unresponsive" or that it's isseu for everyone)

No, that's not what i'm saying at all. Firstly, input-lag is always going to be an issue, the physical nature of the hardware virtually ensures that. But secondly, the caveat to this is that input-lag only becomes an issue when it becomes significantly high. So, hence my repeated question: Where is the evidence that the input-lag introduced by this fix is sufficiently high enough to affect your overall gameplay experience? We have solid evidence of micro-stutter affecting gameplay, but we have zero evidence that the input-lag is sufficiently high enough to cause problems. Clear enough?

I haven't followed FM recently, but on other forums at least people have changed hardware to reduce input lag, they've chosen not to use v-sync due input lag etc, as techreport article says, nVidia microstuttering fix adds up to tens of milliseconds of input lag, which is on level or even more than many monitors add (and monitors are generally the biggest inputlag generator)
Even though I haven't followed FM much, even in this very thread there's at least one user who recognizes input lag as a problem.

Again, see my paragraph above. I accept what you're saying, but where is your evidence that the input-lag significantly contributes to this CB? You make these claims without actually having any evidence to back it up, you just keep saying it must be bad, because input-lag is bad, MmmmKay?

If you only read FM forums, can't say much about it as I have no idea of the general situation here, however what I said above this.

Also, if input lag isn't issue, why are games getting fixes for it? Why are people playing without vsync even though their monitor can't display all the frames rendered? etc etc

[icemanchilled]

Enough BS CB. Give me the solid proof, not anecdotal evidence, which is all you've got so far. You don't know if the lag is sufficient to cause issues, you don't know precisely how much lag there is, you don't know what games, if any, are affected, you don't know if thw graphics subsystem lag acts on the user in the same way monitor lag does, you don't actually know any facts about it. So unless you can present some quantifiable evidence for your claim, you're just spreading unfounded rumours, yet again. Didn't take you long to slip back into your old ways did it? I think i'm done here anyway tbh, i said i don't want to slip back into the old ways, and i don't.

[valtterieranen]

Well let me be one to complain about AFR added input lag. It bothered me with 4870 crossfire and it was immediately apparent with 480 SLI. It's not much of an issue to reaction time but it make the mouse aiming feel floaty.

[kitch9]

I've read somewhere a while that NV deliberately introduce a frames lag when SLI is used to combat micro stutter, where as up to the point of me reading that article AMD did not.

In fact it may have been a test on Hardforum where there came to the conclusion Xfire felt slightly sharper but more stuttery so they said NV's method was best.

Having used SLI and Xfire on a 120hz monitor I can honestly say I'll stick with a single GPU from now on thanks.