FPS Question

[FX-Patrick]

HMMMMMMM...I honestly think people rage because they can't shoot a gun perfectly in a video game. KTHXCOMEAGAIN! :D

[MV-Warren]

Originally Posted by Spartanfrog

There is also the placebo affect. Tell your clanmates its a 1000 FPS server when its only 512 FPS and see what they say :P.

Like i said i did this, i have a 512 fps and its so good the guys think its a 1000fps lol so to me fps isnt really something of importance

[MjrNuT]

I hope this is ok, but I have a couple links to share regarding this "mystery" topic.

Optimizing a Dedicated Server - Steam Support Wiki

oPg Gaming Blog Archive SRCDS Tech: The Rest…

srcds.com forums · Server FPS and Tickrate

To me there are 2 separate issues regarding the role of Server FPS.

First being a stability measurement of the server itself, which some would say translates or means performance (sort of).

Second being the effect of the client(s).

Server side -

FPS is a direct measure of calculating a set number of frames per unit time. This is limited purely by the power of the CPU, clock speed, and related things that would just nerd this up. It would be fair to say that if a server can run at 512, 1k, 2K, etc. at less than 40% usage at 95% confidence level, then the server is stable. This is said explicitly excluding the client phenomenon. Why b/c the aforementioned aspects of the CPU, clock speed, blah blah is dictated by the HARDWARE, OS, applications, etc. (We all know everyone is running different schtuff and some people have noted that and attempted poor comparisons).

So now you have a server that confidently does calculations, b/c that is what these things do, crunch numbers. Those numbers have to be distributed, which is the bandwidth pipe. Think of this as a fire hydrant opened on a hot summer afternoon. It just spews out the gallons of water at the rate dictated by public works, but also the size of that hydrant (and other things not needing mention here).

Now enter the client, which we all know come from all walks of life (i.e., broadband, small band, or w/e). Each client's connection varies so much to travel to the server, we all know this. So here is my quick and dirty analogy to the fire hydrant -- how many hoses can you hook up and still maintain the same output from the hydrant, despite the number and length of each connected hose?

Ideally, you want the hydrant to match those hoses such that everyone is equal, ergo better hig registration or w/e you want to call it, more simply no latency. Nonetheless, there is latency, as the hydrant would have to make to pump more water out to maintain that ideal by turning on a supplemental pump or opening the flood gates on the dam. LOL If it cannot be increased, then the latency will occur.

So then each of those fire engines hooked up to it...they were told that the hydrant would maintain its flow for x hoses, but may in fact ended up w/ some dry lines.

I concur w/ John and some others that client experience varies immensely and that no one can ever really say that one FPS setting is better than another w/o considering the devil in the details -- hardware, network, configuration, bandwidth, on and on...

Apologies for the long windedness. Attempting to make sense of something that is complex, b/c it truly is.

Regards,

[MV-Warren]

Yea i recently went exploring on the valve/steampowered site and found this msg and link as well

WHY HIGHER FPS?
The key reason to run higher FPS is the render time. At 1000FPS, the server is rendering one frame every 1 millisecond (ms). This means that the worst-case adder to the player ping is only 1ms, IE: the player gets more accurate data and can get it more often.

At 300FPS it's only 3ms which is perfectly acceptable, but at 100FPS it's 10ms, which is a significant percentage of a 100 ping (10%). A player with a 100 ping would actually be getting 110ms response time from the server. Many AMD systems will only run 60FPS without the ping booster which is 17ms.

This is still not too significant but it can change the feel and response time of your server for players.

Without the FPS Boost your server will use significantly less CPU but accuracy may suffer.





Optimizing a Dedicated Server - Steam Support Wiki
look at the very bottom of the page.

[Daniel]

Originally Posted by MV-Warren

Yea i recently went exploring on the valve/steampowered site and found this msg and link as well

WHY HIGHER FPS?
The key reason to run higher FPS is the render time. At 1000FPS, the server is rendering one frame every 1 millisecond (ms). This means that the worst-case adder to the player ping is only 1ms, IE: the player gets more accurate data and can get it more often.

At 300FPS it's only 3ms which is perfectly acceptable, but at 100FPS it's 10ms, which is a significant percentage of a 100 ping (10%). A player with a 100 ping would actually be getting 110ms response time from the server. Many AMD systems will only run 60FPS without the ping booster which is 17ms.

This is still not too significant but it can change the feel and response time of your server for players.

Without the FPS Boost your server will use significantly less CPU but accuracy may suffer.





Optimizing a Dedicated Server - Steam Support Wiki
look at the very bottom of the page.

To expand upon that, if you had a 1000FPS 100Tick server you would only be sending 1 out of every 10 frames/world snapshots rendered because you only have it set to 100 tick, so in an essence those 9 snapshots are wasted. But as a result of the higher FPS, each snapshot that is sent is more accurate because it was rendered, at the very most, 1 millisecond before it was sent to the client(s). But on the flip side, the clients are only sending less than 100 updates per second, so what is the server calculating during the other 900 snapshots?

EDIT: To answer my own question: it's probably predicting what the players are doing, because it will not receive all packets at the same time due to end-user latency. So it's predicting and calculating the "world state" as it receives packets from all of the clients, AND correcting any wrong predictions it made once it got the "real" packet.

[RECONGAMER28]

HyperionServers released their 12,000 FPS server... I played on 1 last night and it was the absolute WORST server I had ever played on. My buddy bought it, $99 a month. Gave me rcon to it, it ranges from 11,987 to 93 FPS, yes... 93. Lower then the normal 128FPS (Standard). Reg was horrid, pings were bad, Dallas server I pinged 89 to, same for my lan team, remind you I live in Kansas. The was no such thing as registery to this server.
Also, the server side FPS would never be consistent, Not once, after many tries (20+) did it stay at or above 6,000 FPS after just 2 rcon stats with no one in the server (using HLSW).

*edit* I'd rather play on a 33tick then that.

[Peter]

Rofl @ 12,000 fps lol!!!!!!!!!!!

[stronius]

Originally Posted by Peter

Rofl @ 12,000 fps lol!!!!!!!!!!!

and i thought i was a spamming forum whore.

[PCS|Erik]

Are you serious?


EDIT: I thought you accidentally put a '1' in front of '2,000' as a typo, I was obviously wrong. They are actually offering 12,000FPS servers... Seems like a huge scam to me in my opinion. I would not trust them w/ anything.

[Peter]

My post wasn't spam.