Overwatch Reduce Buffering Bug

Everything about latency. Tips, testing methods, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
open
Posts: 223
Joined: 02 Jul 2017, 20:46

Overwatch Reduce Buffering Bug

Post by open » 03 Oct 2017, 20:38

Reduce buffering has a bug where if you alt tab out it turns off.

So I have been noticing a few streamers toggle reduce buffering off then on at the start of every round. I asked in a chat today why that was and the response I got was that reduce buffering has a bug where it turns off if you alt tab out.

I'm going to get some highly subjective data on this right about now but I'd love to get some participation from the community on this. Any objective data or subjective experiences would be awesome.

open
Posts: 223
Joined: 02 Jul 2017, 20:46

Re: Overwatch Reduce Buffering Bug

Post by open » 03 Oct 2017, 22:52

Honestly its hard to tell but it does seem like it turns off.

The experience I had was that the computer was reacting before my mouse movements. Like it was ahead in time. This is exactly what you would expect if you were conditioned to a slight input lag and then removed it.

daniloberserk
Posts: 37
Joined: 22 Aug 2017, 16:50

Re: Overwatch Reduce Buffering Bug

Post by daniloberserk » 04 Oct 2017, 00:38

Never experienced this myself. There's a very easy way to test this... Limit the game to 30 FPS. Reduce Buffering is VERY noticeable with 30 FPS and you'll easily if it's working or not.

Sparky
Posts: 682
Joined: 15 Jan 2014, 02:29

Re: Overwatch Reduce Buffering Bug

Post by Sparky » 04 Oct 2017, 11:04

daniloberserk wrote:Never experienced this myself. There's a very easy way to test this... Limit the game to 30 FPS. Reduce Buffering is VERY noticeable with 30 FPS and you'll easily if it's working or not.
If you use an in game cap there shouldn't be any difference. If you use a GPU based cap the difference should be very big.

daniloberserk
Posts: 37
Joined: 22 Aug 2017, 16:50

Re: Overwatch Reduce Buffering Bug

Post by daniloberserk » 04 Oct 2017, 14:02

Sparky wrote:
daniloberserk wrote:Never experienced this myself. There's a very easy way to test this... Limit the game to 30 FPS. Reduce Buffering is VERY noticeable with 30 FPS and you'll easily if it's working or not.
If you use an in game cap there shouldn't be any difference. If you use a GPU based cap the difference should be very big.
There's a big difference in overwatch even using the ingame limiter. Not every game engine works the same way.

open
Posts: 223
Joined: 02 Jul 2017, 20:46

Re: Overwatch Reduce Buffering Bug

Post by open » 04 Oct 2017, 15:37

The in game limiter most likely stalls the game based on the last gpu render time and the current cpu sim time and it would happen after the cpu sim but before the mouse input adjustments.

The buffering happens between the rendering thread and the other game threads. Depending on how its coded it could happen in many cases but its most likely to happen when something stalls the gpu (like vsync wait, nvcp limiter, or long gpu render times). This would give the cpu more time to fill up the buffer with data while its waiting for the gpu. Thus the buffers would be more likely to be one frame old. Probably the best way to test would be with high graphics settings, near 100% gpu usage and low fps. If you can get the game to run 30fps due to gpu bottleneck you could perhaps see a diff of 33ms with reduce buffering.

daniloberserk
Posts: 37
Joined: 22 Aug 2017, 16:50

Re: Overwatch Reduce Buffering Bug

Post by daniloberserk » 04 Oct 2017, 19:43

There's a test here: https://www.youtube.com/watch?v=sITJ3V_ ... 9816886012 regarding the "Reduce Buffering" option. It always decreases the average lag in about 1 frame. The higher the FPS, the less noticeable is the effect, this is the reason I'm saying to lock your framerate to 30 and test. There's a VERY noticeable difference (for the majority of gamers of course) and doesn't require any special tools.

I'm pretty sure it's not really necessary to raise the graphics settings in this case. The tests I linked is with low settings.

notthatjesus
Posts: 11
Joined: 24 Apr 2017, 20:48

Re: Overwatch Reduce Buffering Bug

Post by notthatjesus » 08 Oct 2017, 20:47

I became aware of this issue through this reddit post: https://www.reddit.com/r/OverwatchUnive ... set_their/

Some useful info:
Cause alt-tabbing with reduced buffering bugs you game and makes it GPU bounded and other shit. Actually, if you have one dot in FPS-counter you don't need reduce buffering.
One dot - CPU bounded fps, three - GPU one. Check it with uncapped without reduced buffering. I have one, but if turn render 150%, it is three dots. Uncapped 300 fps one dot no reduce buffreing is lowest possible input lag of 3.33 ms

mello
Posts: 251
Joined: 31 Jan 2014, 04:24

Re: Overwatch Reduce Buffering Bug

Post by mello » 09 Oct 2017, 10:43

A lot of contradicting information about this issue.

https://www.reddit.com/r/Competitiveove ... f_you_are/

Until someone actually tests it for all these scenarios (one dot vs three dots, plus alt-tabbing for both) we will not know what the best setting is.

mello
Posts: 251
Joined: 31 Jan 2014, 04:24

Re: Overwatch Reduce Buffering Bug

Post by mello » 16 Oct 2017, 15:50

Well, i can most likely confirm that this bug infact exists... I have never experienced it before (always 1 dot near the fps count) even with alt-tabbing, but recently i have updated nvidia drivers and i often get 3 dots after alt-tabbing and fps are noticeably lower. So it is a bug in nvidia drivers, because my previous drivers (~18 months old) worked just fine.

Post Reply