Breaking the Source engine

[Traveler]

I can go higher than 500fps in CS1.6 but in some servers I will teleport.

What servers are those? - (the ones where you teleport when you go higher than 500 fps).. are those "regular/ordinary" CS 1.6 community servers or Fastcup.net servers and what do you do when the server can't handle more than 500 fps?.. do you just leave and don't want to have anything to do with it or you actually lower it to a "next best value" like 360-400 fps? - (or whatever it can handle while being higher than 100 fps)

I did some testing in CS 1.6 with Nvidia Reflex Analyzer on a 360Hz display and was getting around 4ms system latency (mouse click latency included) at very high FPS (1500+fps).

But, those fps numbers and that latency can not be achieved online, so it's useless.. right?

[Boop]

What servers are those? - (the ones where you teleport when you go higher than 500 fps).. are those "regular/ordinary" CS 1.6 community servers or Fastcup.net servers and what do you do when the server can't handle more than 500 fps?.. do you just leave and don't want to have anything to do with it or you actually lower it to a "next best value" like 360-400 fps? - (or whatever it can handle while being higher than 100 fps)

Test #1 - ~10ms ping with 1000+ FPS and no issues moving around
https://streamable.com/s9mgiu

Test #2 - ~70ms ping with 500FPS and no issues moving around. Once I increase it to 800FPS I start to teleport, and when I increase it to 1000FPS I can no longer move.
https://streamable.com/i1q2oj

Both tests were done on the same server. I used a VPN to increase my ping to show the differences.

If you get low ping to a game server, you can probably use 1000+ FPS and not experience any teleporting. Try out different FPS caps and see what works as it could change when playing on servers with high ping.

But, those fps numbers and that latency can not be achieved online, so it's useless.. right?

I can easily maintain 500FPS online and it's much better than 100FPS

My Nvidia Reflex Analyzer latency results on 360Hz monitor:
Uncapped 1000+fps -> 4ms average system latency
Capped 500fps -> 5ms average system latency
Capped 100fps -> 9ms average system latency

You can see here how there are system latency diminishing returns the higher your framerate is in-game.

[Traveler]

Test #1 - ~10ms ping with 1000+ FPS and no issues moving around
https://streamable.com/s9mgiu

Test #2 - ~70ms ping with 500FPS and no issues moving around. Once I increase it to 800FPS I start to teleport, and when I increase it to 1000FPS I can no longer move.
https://streamable.com/i1q2oj

Ohh, so all this is related to ping then!
The lower the ping, the higher an fps can go.
This is why when we open up CS on an in-game private server with bots and having virtually 1ms internet latency (or LAN setup) we go 1000+ fps with no issue (if hardware supports it of course).
Does this "fps to ping ratio" hold true for CS:GO, CS:S and other games aswell? If yes, then how did forcedreg come up with 1100 fps as a max allowed before rubberbanding/teleporting in CS:GO?
Btw, you were alone on the server.. did that influence the results?

Both tests were done on the same server. I used a VPN to increase my ping to show the differences.

I've read somewhere that this artifical ping increasing stuff with the help of a VPN can be a rather tedious process when used for the goals of fighting lag compensation.
If my ping is on the low end (20ms) and my opponent has a ping of 100ms, often times he will have a competitive advantage over me because of said lag compensation in CS:GO.
So, in this case having higher ping and lower fps is more beneficial than having lower ping and higher fps?

If you get low ping to a game server, you can probably use 1000+ FPS and not experience any teleporting.

Even in CS 1.6?
And how low are we talkin' here?
.. sub 10ms = 1000+ fps with no teleporting?

Try out different FPS caps and see what works as it could change when playing on servers with high ping.

Yeah, this seems to be the way to go.
Does this mean that all this talk about "cap your frames all the time at this number or that number" isn't always the right thing because of another variable too - monitor's refresh rate?

I can easily maintain 500FPS online and it's much better than 100FPS

But you couldn't maintain 1000+ fps online because you were never able to bring your ping to less than 10ms?

My Nvidia Reflex Analyzer latency results on 360Hz monitor:
Uncapped 1000+fps -> 4ms average system latency
Capped 500fps -> 5ms average system latency
Capped 100fps -> 9ms average system latency

You can see the system latency diminishing returns there.

Why is 4ms the best average you could achieve? - (not saying that's bad or high.. just to be clear ).. is it because of your 360hz monitor limitation?
I see that in your case it's pointless to go above 500 fps in CS 1.6 then, because no competitive advantage between 500 and 1000.
Thank you for testing it like this.
Last question - is Nvidia's Reflex Analyzer best "tool" for tests like these or is it just the most user friendly?

[Boop]

Does this "fps to ping ratio" hold true for CS:GO, CS:S and other games aswell? If yes, then how did forcedreg come up with 1100 fps as a max allowed before rubberbanding/teleporting in CS:GO?
Btw, you were alone on the server.. did that influence the results?

The results seem the same with other players in the server. I haven't tested CS:GO or CS:S but it could be the same.

I've read somewhere that this artifical ping increasing stuff with the help of a VPN can be a rather tedious process when used for the goals of fighting lag compensation.
If my ping is on the low end (20ms) and my opponent has a ping of 100ms, often times he will have a competitive advantage over me because of said lag compensation in CS:GO.
So, in this case having higher ping and lower fps is more beneficial than having lower ping and higher fps?

Lower system latency, lower ping is going to be better

Watch this:

phpBB [video]

But you couldn't maintain 1000+ fps online because you were never able to bring your ping to less than 10ms?

If I connect to a server with higher ping I have to adjust my FPS cap. Seems like 500FPS is good enough of a cap to not have to worry about adjusting when I may switch servers. This applies to CS 1.6 and HLDM for me.

Why is 4ms the best average you could achieve? - (not saying that's bad or high.. just to be clear ).. is it because of your 360hz monitor limitation?
Last question - is Nvidia's Reflex Analyzer best "tool" for tests like these or is it just the most user friendly?

Probably. I'm waiting for the Asus 540Hz monitor to be released.

Reflex Analyzer is pretty easy to use but I'm sure there are other methods that exist. Check out the Input Lag category for DIY latency testers: viewforum.php?f=10

[Traveler]

Lower system latency, lower ping is going to be better

Watch this:

phpBB [video]

That's R6S, I don't think it applies to CS:GO and its high lag compensation.
There's a huge problem in CS:GO when fighting players with high pings while also Valorant for example doesn't have such a big lag compensation built in.
Benefits of low system latency of course apply to every game.

Another thing..
Would 1000 fps cap + 1000hz be an absolute end game when we talk about competitive advantage?
Right now we can do 1000 fps + 360hz, soon it will be 1000 fps + 540hz, but then we have screen tearing, right? Because fps is higher than monitor's refresh rate.
Now, is screen tearing just a visual annoyance or an actual competitive disadvantage?

[Boop]

Now, is screen tearing just a visual annoyance or an actual competitive disadvantage?

Screen tearing at framerates above your refresh rate gives you an advantage. See this article https://blurbusters.com/faq/benefits-of ... resh-rate/

[Traveler]

Screen tearing at framerates above your refresh rate gives you an advantage. See this article https://blurbusters.com/faq/benefits-of ... resh-rate/

Good read!

I got couple questions tho..
Firstly, it says "...visibility of stutters and tearing gradually reduces on average, the higher your frame rate goes above refresh rate."
On a 60Hz monitor + 1000 fps the frames there are a lot lot higher than refresh rate.
While, on a 540Hz monitor + 1000 fps, frames would still be higher than refresh rate, but the difference between Hz and FPS wouldn't be as big as with 60Hz/1000FPS combo.
What does that do to the image on screen then?

Secondly, what is this video exactly? - https://youtu.be/6fiDiRBvuVI

TL;DR of that video is 1024x768 res/400 fps default game engine lock = lowest latency - which then = more competitive advantage?
What I took from this is, based on his video - uncapped fps on a system that can reach up to 904 fps is inferior to that 400 fps cap setup.
He calls 400 fps most stable and says there's no point in going above that.

Plus, how accurate could such testing be if he did that only while stationary and shooting at a wall close proximity?

I watched the Battle(non)sense latest video about Nvidia's RLA and CS:GO is not on the list of supported games, so RLA cannot be used to test CS:GO while moving/running around, only stationary shooting tests.

Then, how does one even properly test CS:GO for latency?

[Traveler]

Also, does this all mean that on LAN/in-game server with bots, there's unlimited frames per second a game (engine) can handle since there's no internet latency from the server to account for?.. and by "handle" I mean no rubberbanding/teleporting or other "game breaking" occurrences once we go past 1000 fps.

[Slender]

Not really answering your question here but just adding to the thread.

Seems like there's quite a few older games that have issues with really high framerates when playing online. Unreal Tournament 2004, CS 1.6/HLDM, Quake 3, etc. The rubberbanding/teleporting basically makes it unplayable. If you create a LAN server, you don't teleport around at high fps, so it's really related to the game server not keeping up with the client.

For CS 1.6, I have a keybind that caps my FPS at 100 with my knife out, and then at 500fps when I have a gun out. I don't experience any teleporting issues and movement works fine.

I guess CS:S would then have the same fps threshold like CS:GO (around 1100 fps) before it "breaks" on community servers.
So it's not game engine related, it's server related.

For a while now UT2004 or any UT for that matter can't be played online in multiplayer mode anyways (EPIC games doing heavy drugs again)

You're telling me you're useing a script to switch between fps numbers in CS 1.6 only to avoid glitches and rubberbanding?

Also, what is the reason developers choose to make a default cap at a certain number (CS 1.6 is only 100 fps) when that is not always the best for achieving competitive advantage in every game?

steam css feel worse on framerate over 400fps, clientmod (v34) work normal with 999 framerate. Maybe limiter problem.

[Traveler]

steam css feel worse on framerate over 400fps, clientmod (v34) work normal with 999 framerate. Maybe limiter problem.

Wow, so CS:S Steam version has a lower fps tolerance than CS 1.6
By "feels worse" what do you mean exactly? Is it the rubberbanding/teleporting like discuseed here earlier?