CS:GO -- Does 2 milliseconds matter?

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.
User avatar
RealNC
Site Admin
Posts: 3730
Joined: 24 Dec 2013, 18:32
Contact:

CS:GO -- Does 2 milliseconds matter?

Post by RealNC » 13 Dec 2014, 00:53

[Moderator's note from Chief Blur Buster -- this was originally part of a different thread, but a worthy topic meritorious of its very own thread!]
Chief Blur Buster wrote:On the other hand -- If you are playing on a GTX Titan, and just 4X AA (or less), you can easily get 150fps+ with single core rendering.
I'm on a GTX 780, which isn't *that* much slower compared to a Titan. The problem is that the 150-ish FPS it gives with single-core rendering is not useful, since it's a best-case. The minimum FPS can drop down to 60 in many situations, and the GPU isn't even getting used at all when that happens. In those situations, the bottleneck is the CPU, which seems to be the reason why enabling multicore rendering gives such a huge performance boost in this game. I can drop the graphics quality settings down to minimum and it doesn't help a bit in those situations.

This game is quite light on GPU usage (most maps are very well hand-optimized, using manually placed visleafs, where stuff that you don't see isn't even considered for rendering,) which means that uncapped frame rates will be CPU bottlenecked at some point, especially since the network engine also eats up quite a bit of CPU.

If you fall below 128FPS, your hit registration is going to suffer, unless you're playing on servers with a tick rate of 64. This is much, much worse than multicore rendering's slightly higher input lag. Consider that an AWP flick-shot is in the ballpark of 2-3ms. This is a muscle-memory reflex, where times are really that low. In many occasions, 60FPS Twitch.tv broadcasts of top-level tournaments are not even able to capture these shots.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

flood
Posts: 929
Joined: 21 Dec 2013, 01:25

Re: Does 2 milliseconds matter?

Post by flood » 15 Dec 2014, 05:24

RealNC wrote:Consider that an AWP flick-shot is in the ballpark of 2-3ms.
uh no

User avatar
RealNC
Site Admin
Posts: 3730
Joined: 24 Dec 2013, 18:32
Contact:

Re: Does 2 milliseconds matter?

Post by RealNC » 15 Dec 2014, 07:41

flood wrote:
RealNC wrote:Consider that an AWP flick-shot is in the ballpark of 2-3ms.
uh no
Oh yes. The moment between you moving the mouse and clicking the button can be that low. Note: not the reaction time of you seeing the player and moving the mouse; that's in the 130ms range at best. I mean time between mouse move and click. It can be almost instantaneous; you pretty much click the button at the same time as you move the mouse. If your FPS is below tickrate, the hit might not be registered at the correct position.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Does 2 milliseconds matter?

Post by Chief Blur Buster » 15 Dec 2014, 08:54

RealNC wrote:
flood wrote:
RealNC wrote:Consider that an AWP flick-shot is in the ballpark of 2-3ms.
uh no
Oh yes. The moment between you moving the mouse and clicking the button can be that low. Note: not the reaction time of you seeing the player and moving the mouse; that's in the 130ms range at best. I mean time between mouse move and click. It can be almost instantaneous; you pretty much click the button at the same time as you move the mouse. If your FPS is below tickrate, the hit might not be registered at the correct position.
Interesting observations.

Based on tests in "other areas", for highly attuned atheletic situations: 2-3ms doesn't seem unrealistic because it's a "react on the starting pistol" situation. Tests have shown that some olympic racers can consistently react LESS than 10ms "variability" from a starting pistol.
Olympic race #1, person starts running at 132ms after the shot
Olympic race #2, person starts running at 137ms after the shot
Olympic race #3, person starts running at 133ms after the shot

Likewise, the mouse click and movement can be so precisely co-ordinated (synchronized) in certain types of situations, so I have totally no doubt that 2-3ms is probably accurate for Event "A" synchronized to Event "B" if you've had lots of practice precisely timing the two events (e.g. mouse button and mouse movement). You don't even need Tiger Woods precision to achieve less than 10ms synchronization between two actions. Although, the skeptic in me would appreciate, high speed camera tests on the "2-3ms" territory, but its well within the realm of realisticness due to the starting-pistol-variance test.

Although I don't compete in the professional leagues, I've been around long enough not to disagree: I've made observation that milliseconds DO matter.

Human #1 shoots at average ~160ms reaction time
Human #2 shoots at average ~154ms reaction time
Human #3 shoots at average ~155ms reaction time.
-- Reaction time differences where milliseconds of advantage (human-wise or equipment-wise) can matter

But since you say it's not reaction time differences, it'd be interesting to see high speed camera tests. This is an interesting new area of scientific study for high speed cameras.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

badugib
Posts: 7
Joined: 02 Jun 2014, 10:15

Re: Does 2 milliseconds matter?

Post by badugib » 15 Dec 2014, 13:31

Maybe flood should do these tests lol. Since he's already doing various counter-strike input lag tests anyways.

spacediver
Posts: 505
Joined: 18 Dec 2013, 23:51

Re: Does 2 milliseconds matter?

Post by spacediver » 15 Dec 2014, 18:06

that's a really interesting question, and it clearly depends on distance that mouse is moved and the mechanical properties of the mouse button. Should be easy to test with flood's set up (tho he'd use a real mouse, not the arduino mouse). Have the screen start off with black frames, and upon mouse movement, render a single white frame, followed by black frames, and upon subsequent button press, render another single white frame. Assuming that the photodiode is placed such that it can register a single tear line at any position on the screen, it should be able to register both events.

Answer = time between photodiode events.

at 9000 fps (which is what his system is capable of), precision of display events will be about 0.1 ms. But because the two mouse events (movement and button press) are independently polled, the precision will be limited by 1 USB frame (1 ms).

Test using a very brief and fast flick shot.

flood
Posts: 929
Joined: 21 Dec 2013, 01:25

Re: Does 2 milliseconds matter?

Post by flood » 15 Dec 2014, 20:46

time between mouse move and click is definitely at least 30ms. probably ~100ms for fast flicks

starting from rest you cannot move a mouse a significant amount in 2-3ms.

for flick shots its really more about fast aim than timing. timing is relevant for reaction shots where you're holding angle and planning to shoot as soon as you see anything. for that, 10ms is a relevant number but still only marginally

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: CS:GO -- Does 2 milliseconds matter?

Post by Chief Blur Buster » 15 Dec 2014, 22:32

flood wrote:time between mouse move and click is definitely at least 30ms. probably ~100ms for fast flicks

starting from rest you cannot move a mouse a significant amount in 2-3ms.
Oh, I didn't think RealNC meant that, from BEGINNING of flick to the shot. At least I don't think....
Otherwise, I would agree with you! That would simply be superhuman.

I interpreted it differently:
-- It meant timing the click of mouse within +/- 2-3ms of intended mouse button press timing.
For event such as:
...Shooting right after finishing the flick (timing of END of flick and the PRESS of mouse button).
...Or mousing past an enemy while shooting, or shooting moving targets, sniping, etc. Shooting without stopping moving mouse, requires a lot of precision, especially if you're mousing fast past far-away enemies, sniping as many frags as possible in a short time period. Say, you're mousing stuff past your crosshairs at 1000 pixels/second without stopping moving the mouse, and the faraway enemy is only 6 pixels wide, you need to press the mouse button at an intended time range of 6ms, for that shot to land on the enemy. That would be +/- 3ms of dead center. Since some pro players are able to shoot without stopping moving the mouse to aim, this may also be what is meant too.

This might be what RealNC meant, operations that require only 2-3ms error variances...

If you interpret it this way, it makes a lot of sense.

RealNC, can you clarify?

(Gee, I wonder if Blur Busters needs a CS:GO specific forum, or Competitive Gaming forum)
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: CS:GO -- Does 2 milliseconds matter?

Post by Chief Blur Buster » 15 Dec 2014, 22:55

[This topic has been split into a new thread]
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

flood
Posts: 929
Joined: 21 Dec 2013, 01:25

Re: CS:GO -- Does 2 milliseconds matter?

Post by flood » 16 Dec 2014, 02:21

lets just say that no one is able to time stuff in csgo within 3ms consistently.

Post Reply