Page 7 of 65

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 14:55
by kevindd992002
jorimt wrote:
kevindd992002 wrote:Ok, that makes sense.

I just re-read the optimal settings page and had a better sense of what you're saying. I have a question regarding FPS competitive games like CS:GO and Overwatch (especially CS:GO) though. There are mixed reviews all over the Internet regarding using GSYNC with CS:GO vs. turning it off and just using ULMB with as high as possible framerate that it can go. I know there will be tearing but most say that super high framerates tend to reduce tearing and makes FPS games more "smoother" to the feel. What is your take on this?
First off, what is your max refresh rate, and how how high can you sustain framerate above the refresh rate in CS:GO?
I'm using a 1080Ti at 1080p so I can get 250-300 fps in CSGO. And I'm using a 144Hz GSYNC monitor.
Chief Blur Buster wrote:For low-lag Blur Reduction, 1000fps VSYNC OFF is indeed a good compromise with ULMB since strobed variable refresh rate isn't currently yet mainstream. If the framerate-locked option is not available (due to VSYNC ON input lag concerns) then massively exceed the framerate to minimize the microstutter/tearing disadvantage of VSYNC OFF -- the insane framerates compensates a lot.

We'll be doing ultra-detailed lag testing with ULMB (and other strobe backlights) in the coming months too.
So are you saying that anything below 1000fps is not worth not using GSYNC?

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 14:59
by Chief Blur Buster
kevindd992002 wrote:So are you saying that anything below 1000fps is not worth not using GSYNC?
Not necessarily. The threshold of "not worth it" depends. And it is also a personal preference too.

That said, 1000fps+ is so far beyond the max Hz of current VRR monitors, and lag-reduction so consistent, that it's reliably lower lag than GSYNC (until 1000Hz GSYNC monitors arrive).

The scanout lag affects lag measurement methodology ("lag to first screen reaction anywhere" versus "lag for center of screen") and will affect this greatly. The min/max/average will behave differently for different lag measurement methodologies.

For screen centre (measuring lag only at crosshairs), the mathematical crossover point where VSYNC OFF starts to consistently have lower average input lag than a well-optimized GSYNC system -- is roughly twice the framerate of refresh rate. But if you're measuring to first screen reaction anywhere, the math suddenly changes. This is relevant for close-quarters combat, peripheral vision, screen flashes, explosions, etc -- where you can react to frameslices (VSYNC OFF) delivered sooner in previous refresh cycle. If most of what you do is only visible in screen centre, lag measurement methodology can change all over again (due to scanout lag factor). Framerates not much above GSYNC max rate doesn't always consistently favour VSYNC OFF, but ultra-high framerates clearly show lag advantage.

Although for this specific lag test run, we've been measuring "to first screen reaction, anywhere", the Average bar (lime green color) will roughly (mathematically) generally correspond to screen-centre lag (crosshairs) -- the lag measurement methodology does not map 1:1 -- but it illustrates the lag-reducing benefits of frame rates far exceeding monitor's refresh rate -- with diminishing returns.

Very clearly, the worst-case "VSYNC OFF" lag consistently keeps going down, the higher the framerate you go (even when you've already maxed-out the best case lag). You notice when you reach quadruple-digit frame rates, the worst-case VSYNC OFF lag (dark grey bar) start to beat a lot of "average lag" GSYNC numbers (lime green bar) of properly-configured low-lag GSYNC.

240Hz version:
Image

144Hz version:
Image

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 15:04
by Glide
jorimt wrote:
Glide wrote:1. I had been testing with G-Sync disabled, since I was only checking for tearing in Borderless Windowed Mode.
Glide wrote:Right, but if you enable V-Sync in Dishonored 2, you can break past the 120 FPS limit.
I can replicate the in-game V-SYNC quirk with G-SYNC, but I can't replicate the tearing you are seeing with borderless or windowed mode in Dishonored 2.

I tried all combinations of Game Mode and fullscreen optimization settings with G-SYNC and in-game V-SYNC disabled, even at 60Hz with frames well over 100, and not a hint of tearing to be seen. There was loads of microstutter, but that's to be expected with the DWM triple buffer solution.

It must be system specific. Are you sure you haven't been on a slightly earlier version of the Creators update since it released? For reference, mine is currently Windows 10 Home version 1703, build 15063.413. I'm thinking they may have corrected that behavior in a slightly later release, and that's why I didn't experience it while testing.

If that isn't it (likely it isn't), have you tried deleting your config file, and trying to see if you can replicate the tearing in borderless again?
Version 1703, build 15063.413 here too. I waited for the final release, used the Microsoft tool to create a boot drive, and did a clean install. It was a new PC build anyway.
It's a Pro license rather than Home, but that shouldn't make a difference.

The only thing I can think that it could possibly be is if you don't have a WDDM 2.2 video driver installed?
I don't think that should affect it, since the feature was there for DX12 UWP games prior to the Creator's Update being released.

It just seems like the full-screen optimizations are not working or being enabled on your system for some reason.
I've even tried adding another display (I normally only use a single display) and it's still allowed to tear in Borderless Windowed Mode if V-Sync is disabled.
Windowed Mode will never tear, only Borderless (fullscreen) Windowed Mode, and only in the games white-listed to use the fullscreen optimizations.

I deleted my config file for Dishonored 2 and it didn't make a difference. As I said, I've tested maybe 20 games off that list of "fullscreen optimized" games now.
Chief Blur Buster wrote:I think what you probably meant:
It was, and I've edited my post now to reflect that.

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 15:23
by jorimt
kevindd992002 wrote: I'm using a 1080Ti at 1080p so I can get 250-300 fps in CSGO. And I'm using a 144Hz GSYNC monitor.
Well good news is, you can get much, much higher framerates with that setup at 1080p if you set the game's "fps_max" setting to a higher limit. Try a Steam launch option of +fps_max 999 for the game.

Let me know how many frames you get in an average match with that (G-SYNC & V-SYNC disabled), and I can provide further info on what input lag reductions you can expect to be having over G-SYNC with a -3 FPS limit.
Glide wrote: Version 1703, build 15063.413 here too.
Then I'm stumped for now. I was on the latest 64-bit driver at the time of testing, and I'm on the latest driver (382.53) as of now :!:

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 16:27
by Chief Blur Buster
It appears dwm.exe (Desktop Window Manager) is an extremely picky beast about permitting true VSYNC OFF with tearing in (borderless) windowed mode. It's very brand new, being a Creator's Update feature.

For some computer systems, due to some unknown un-met prerequisite, it seems almost easier to break into Fort Knox than to convince Windows to give you the "VSYNC-OFF-with-tearing" borderless windowed mode. Yet other systems let it happily happen.

If any forum member knows why some systems isn't letting it happen, we'd like to know. I bet Jorim would be interested in running a quick input lag run to compare to exclusive full screen VSYNC OFF mode. Barring that, I might have to give a computer a "clean Creator Update install" treatment.

Glide, does the monitor say it is in GSYNC mode or fixed-Hz mode? Does borderless windowed tearing happen for you in either mode for you?

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 17:49
by Glide
Chief Blur Buster wrote:Glide, does the monitor say it is in GSYNC mode or fixed-Hz mode? Does borderless windowed tearing happen for you in either mode for you?
G-Sync will activate, even if G-Sync is set to Fullscreen Only mode, not Fullscreen + Windowed.
Tearing only happens with the fullscreen optimizations enabled, and in supported games.
If I disable them, or play a game not on the list, it won't tear. (triple-buffering as usual)
If I use windowed mode, rather than borderless windowed mode, it won't tear either.

I wish I had a better idea of what the conditions are that allow/prevent it from activating. I didn't have to do anything for it to work.

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 19:40
by kevindd992002
@Chief

Thanks, that makes total sense.

@jorimt

Yeah, I figured I'm getting the same framerates when I was using a 1070. Fps_max is set at 999 even before you mentioned it. I'm still using a 2600K CPU so I think this is a case of a CPU bottlenecking the GPU.

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 19:53
by RealNC
Most multiplayer games, including CS:GO, are doing a lot of stuff on the CPU, which is why they are all CPU-limited. CS:GO especially is rather light on the GPU. It can run at 400FPS even with something like a 970 or 1060. It's the CPU that matters most.

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 21 Jun 2017, 19:58
by jorimt
kevindd992002 wrote:Yeah, I figured I'm getting the same framerates when I was using a 1070. Fps_max is set at 999 even before you mentioned it. I'm still using a 2600K CPU so I think this is a case of a CPU bottlenecking the GPU.
Ah, that explains it. You omitted the CPU in your initial post, so I assumed you had a more modern CPU to go along with that brand new GPU.

I only have a 1080, but I do have a i7-4770k overclocked to 4.2GHz, and I can easily achieve 600+ framerates with V-SYNC off in an online CS:GO match.

You should consider a CPU upgrade, as your current CPU will seriously bottleneck your GPU in any game. Even an upgrade to my CPU could see your 1080 Ti pushing up to twice the framerates (probably more) most of your games get now.

Knowing this, if you can only achieve 300 FPS on a 144Hz G-SYNC monitor, if we count middle screen (crosshair-level) updates only, you'd be getting an input latency reduction of roughly 1-3ms with V-SYNC OFF over G-SYNC + V-SYNC with a -3 FPS in-game limit.

Whether that's worth it is up to you, and is down to situation and preference; tearing and stutter-free consistency vs. slightly more responsive input w/tearing and microstutter. Whether you use ULMB would also factor in.

Re: Blur Buster's G-SYNC 101 Series Discussion

Posted: 22 Jun 2017, 15:55
by sekta
Did you test ingame limiter with RTSS limiter active at the same time?
Would be interested to see when both are set to the same value, or ingame at Hz - 3 and RTSS at Hz - 2, or some other staggered set up depending on how much the ingame limiter fluctuates.