Blur Busters Forums

Hello,
some game developers at the 3DCenter forum claim that there is probably an additional input lag with the new drivers due to tiled resources:
http://www.forum-3dcenter.org/vbulletin ... st10171896

Most users can't check this, at least not objectively.
blurbusters.com made input lag measurements with G-Sync. It would be great if you could do the same with the new driver, compared to the previous one (337 vs 335).

I'd suggest Windows 8.1, a 4770k CPU and Battlefield 4 (empty multiplayer server) for checking. Ideally, there should be a test regarding the CPU limit (low resolution or 25% resolution scale) and also the GPU limit (1440p + 4xMSAA).

Any interest in this?

Thanks!

aufkrawall wrote:Any interest in this?

I definitely have an interest.
That said, there is a definite backlog of articles I need to write! So it may be a long while for this specific test.
(Got freelance reporting/testing experience? Send me a PM!)

I would test BF4 locally only to prevent network latencies from figuring into the calculations, as graphics drivers normally wouldn't affect network gameplay. Reduce the number of variables that almost definitively wouldn't affect things, to try and isolate the driver behaviors.

There's also this interesting thread
http://www.overclock.net/t/1388836/inpu ... ia-drivers

For this type of stuff, a lot of it may just be in the head and I'd take any subjective claims with a grain of salt.

flood wrote:There's also this interesting thread
http://www.overclock.net/t/1388836/inpu ... ia-drivers

I would like to point out that this guy is talking out of his ass. Where's the proof of those extraordinary claims?

Well, you never know -- some competitive players are really good at feeling a 1 frame change in input lag occuring.

However, the lag can behave differently in windowed mode (Windows compositing) versus full screen mode. There could be cases where input lag changed during windowed mode, but that full screen mode was unaffected, so these modes kind of need to be measured differently.

Chief Blur Buster wrote: I definitely have an interest.
That said, there is a definite backlog of articles I need to write! So it may be a long while for this specific test.
(Got freelance reporting/testing experience? Send me a PM!)
,,,,,,,8

I got a Qnix with bypass panel, or do you think a CRT is required?
If I had the equipment I'd definitely try serious measurements.

Chief Blur Buster wrote: I would test BF4 locally only to prevent network latencies from figuring into the calculations, as graphics drivers normally wouldn't affect network gameplay. Reduce the number of variables that almost definitively wouldn't affect things, to try and isolate the driver behaviors.

I agree, SP test would be more reliable.

Input lag tests AMD vs. Nvidia would be fundamental research in general, I don't think this has ever been checked.

I'm interested in this as well. I'm on the new drivers.

r0ach from OCN also made a quite comprehensive optimization guide for input lag. http://www.overclock.net/t/1433882/gami ... ch-edition

But he doesn't provide any actual tests. Would be nice if someone could actually test all these things.

Haste wrote:r0ach from OCN also made a quite comprehensive optimization guide for input lag. http://www.overclock.net/t/1433882/gami ... ch-edition

But he doesn't provide any actual tests. Would be nice if someone could actually test all these things.

I laughed so hard at his claim that RAM CAS latency settings affect input lag. These latencies are in nanoseconds. This guy will be getting close to being able to observe quantum effects with the naked eye very soon now. I think he found the Higgs boson way before CERN because his mouse felt heavier due to its mass increasing at 1000Hz.

Pretty much most of this "tweaks" are at least funny ("On-board audio - Disable. Enormous input lag", "Legacy USB support - mouse feels like crap with it enabled".)

This is placebo at its best.

RealNC wrote:

Haste wrote:r0ach from OCN also made a quite comprehensive optimization guide for input lag. http://www.overclock.net/t/1433882/gami ... ch-edition

But he doesn't provide any actual tests. Would be nice if someone could actually test all these things.

I laughed so hard at his claim that RAM CAS latency settings affect input lag. These latencies are in nanoseconds. This guy will be getting close to being able to observe quantum effects with the naked eye very soon now. I think he found the Higgs boson way before CERN because his mouse felt heavier due to its mass increasing at 1000Hz.

Pretty much most of this "tweaks" are at least funny ("On-board audio - Disable. Enormous input lag", "Legacy USB support - mouse feels like crap with it enabled".)

This is placebo at its best.

Agreed. it is either placebo or this guy's system is seriously messed up. Most of that stuff doesn't affect anything.

RealNC wrote: I laughed so hard at his claim that RAM CAS latency settings affect input lag. These latencies are in nanoseconds.

yea that's probably all in his head...

still it'd be interesting to see how driver versions/configuration affect input lag.

Chief Blur Buster wrote:Well, you never know -- some competitive players are really good at feeling a 1 frame change in input lag occuring.

yea 1 60hz frame is definitely possible to feel. on the other hand, after one is accustomed to an additional ~10ms of lag, it isn't much of a problem.

but I must say I've never noticed significant input lag related to graphics driver versions. The only things that cause noticeable input lag are vsync, and aero compositing for windowed applications.

random note: I always feel typing in firefox on linux is more responsive than on windows, even with aero disabled. wonder if it's just in my head or if there is really a difference