90% finished input lag in CS2/Valorant - a datapoint

[joseph_from_pilsen]

Here someone measured CS2 with some hard data... Sadly not enough settings, but still better than nothing.

https://github.com/Szwagi/cs2-input-lag-analysis

Its raw data, persisted input lag is always higher at any settings, because the effect appears with NEXT tick (1/64s = up to 14ms later)

[F1zus]

All CS2 tests are performed on DX11. This is trash, bullshit.
DX11 provides better performance but causes greater input lag and higher frame times. Vsync ON +6ms on DX11.
In a CS2 game you only need to use Vulkan. It provides the least input latency and does not increase it when vsync is enabled (+0ms)
I don’t know, maybe these are the values on my AMD video card. As we know, AMD video cards are most efficient with Vulkan.

[The90sPope]

That's a bald statement. Why Vulkan? Have you run any tests to back your claim?

[franisagod]

That's a bald statement. Why Vulkan? Have you run any tests to back your claim?

vulkan give me best latency but poor performance compared to dx11 // test on nvidia card

[Enigma1]

Here to provide a data point and hopefully give something back to the Blurbuster community. Everything in this post is subjective and based on my own experiences/testing.
[long reply]

So with none do you mean to leave it at default which is 26 hex or 38 in decimal?

[samzor]

dear yee

Fps max 0 vs 300-9999: medium in cs2, none in csgo

I don't get what you want to say in your post. You want to say that doing fps_max 0 is worse in term of input lag than capping the fps_max 240 for a 240HZ monitor for example ?

[yee]

Hi samzor,

Apologies for the confusion - for the settings I'm running, I find any fps_max values other than zero adds minute stutters to my shots. Almost as if the mouse skips a beat in a gunfight. Doesn't really happen when people aren't shooting at you.

Now this might have been changed/patched after the recent big update but actually reasonaly recently I got a couple friends to try out fps_max 0 and they seem to agree that fps_max 0 is preferable at the cost of higher fan noise and temps.

Hopefully this is helpful.

[xShino]

So basically you are saying getting more FPS is better, if I may explain that in other words.

Or do you think, capping fps_max to 200 is bad if you get 150 FPS max anyway?

[yee]

I secretly feel like fps_max 1000 means every millisecond check if previous frame has finished. If it did, delay next frame till next check.

So yes, I do think fps_max 0 is superior for me. That being said, my constraints (bottlenecks on the systems performance) may not be the same as yours so I suggest you run a variation of those settings and play a few games on each. My suspicion is that at 200 fps they won't make a discernable difference - the source 2 engine is well optimized to run at these speeds.

As an anecdote, I recall someone mentioning a quote from a valve employee saying that their jobs would be 100x easier if they locked players to a certain fps.

Good luck!

[samzor]

I never tried to caps my fps so far, so I don't know to be honest.
My current setup is:

AMD 7800X3D
Asus 3070
32GB DDR5 Corsair 6000MHZ
XL2546K