FPS when testing?


Dustin Diamond
05-16-2006, 02:45 PM
Hey y'all. What kind of FPS should i be shooting for when I run through my map. I know this is kinda vague, but I want to address any performance issues before my map gets too involved.

Essentially, is there any coorelation between FPS when I play the map with just one person on LAN mode and tons of people on a server? I know there are tons of factors, but just ballpark it. Currently the lowest I get in certain areas is 65ish, running an AMD 3500, 6800gt, and a gig-o-ram (1280x1024, video options maxed, 2xAA). I'm curious as to whether those places will lag in full multiplayer. Maybe there's a percentage drop you guys usually notice.

There might not even be an answer to this question due to the vagueness of my details, but I figured I'd toss it out to the minds of the dod world.

[SAS]==Dirty_Harry
05-16-2006, 03:35 PM
Different people have different views on this due to the spec of the individuals pc.

The only way I have found of getting a true idea is to run one of the standard maps on a LAN and compare the FPS you get on that with your spec, to the FPS you are getting in your map.

Hope this helps

Dustin Diamond
05-16-2006, 03:48 PM
Originally posted by [SAS]==Dirty_Harry
Different people have different views on this due to the spec of the individuals pc.

The only way I have found of getting a true idea is to run one of the standard maps on a LAN and compare the FPS you get on that with your spec, to the FPS you are getting in your map.

Hope this helps

Oh great idea. I didn't even think of that. Thanks!

Furyo
05-16-2006, 04:00 PM
That would only be a "true" way if the original maps were as optimized as can be, and they're not. And that doesn't take into account the biggest factor: server load.

The true way of knowing if you have good enough FPS is to get a test set up on a full 32 men server. If anyone complains, it's not good enough

MrGrubby
05-16-2006, 06:30 PM
To get an idea about what types of FPS you should be getting while you're alone in a LAN, load up other maps (the stock maps in particular) and see what their FPS is. You'd want to aim for that value or higher while in your own map. You should look for higher FPS in open area, due to the likelyhood of having to render more player models than normal during server play.

Ca-Chicken-Soup
05-16-2006, 06:32 PM
Use +showbudget (http://developer.valvesoftware.com/wiki/Showbudget) and in the top right-hand corner there will be a fps indicator of 3D something-a-rather. I think this is kind of a general fps dictator showing how it will run on other computers, keep it above 40.

Or you could run an offical map, get the average and see how your map compares, if it's lower then try to make it better

Someth|ngW|cked
05-16-2006, 07:42 PM
Yeah just compare your fps to official maps, thats the best way

JakeParlay
05-16-2006, 08:27 PM
with a bunch of players on the map, i think it creates stress in 2 ways: network load, and rendering load. the engine must negotiate all the net traffic, but also draw all the additional players, muzzle flashes, smoke grenades, etc.

what I like to do, just to get a ballpark sense of FPS:

load the map, join axis, then -

sv_cheats 1
mp_limitteams 20
bot (8 times)

let the bots run around till theyre all spread out, then hit

bot_mimic 1

every time you click, all the bots will fire . It's not a great test, but it does help. It will jack up the FPS hit a lot more than you running around solo

Agent Smith
05-22-2006, 08:26 PM
Of course with the addition of HDR lighting it introduces another problem, should the fps be measured with HDR on (which can drop anywhere from 20-50 frames off a maps framerate) or without.

Generally for my maps if the framerate is good without HDR (60+ fps) then I'm happy with it. While I can run HDR on my PC, I expect a major performance drop when running HDR. However I don't expect one when not running HDR, so if the frames are good without HDR then that should be ok.

HDR is really only for those with beefy machines, or those who are willing to have that drop in performance.

Neutrino
05-22-2006, 09:37 PM
Originally posted by Someth|ngW|cked
Yeah just compare your fps to official maps, thats the best way

and go for a higher standard of performance. dont risk too much fps just to impress with visuals. yes, have something that looks nice, but dont go overboard.

personally, i like the way the ealier cs:source maps were made (dust, office..ect) great improvement of visuals as well as amazing performance for the engine.

:cool:

nave
05-23-2006, 11:08 AM
Relative fps checking isn't the best way. It is the only way.

Unless you just rely on feedback from testers I guess.

[SAS]==Colster==
05-23-2006, 01:56 PM
Bear in mind on a fast vis compile you will generally see lower fps than when you do a full compile, can gain up to 20 through a full compile but don't rely on it

Neutrino
05-23-2006, 06:54 PM
^^yeah

ultimately. your best testing method for FPS is with a group of people. that takes into account for the player models, and when action actually takes place. so make sure to do that :cool:

Day of Defeat Forum Archive created by Neil Jedrzejewski.

This in an partial archive of the old Day of Defeat forums orignally hosted by Valve Software LLC.
Material has been archived for the purpose of creating a knowledge base from messages posted between 2003 and 2008.