PDA

View Full Version : Reasonable FPS?


jsera
11-13-2003, 01:31 PM
What is a reasonable FPS for a map?

I have a PII 800 w/ 512 mb and a 256mb nvidia card.

Some of the user made maps I have played have 10 FPS in some places.

I have tried to keep my maps above 30 FPS.

What is a reasonable FPS for a level?
What about r_speeds? Anyone use them anymore?

Thanks fopr the input.
J.Sera

wedge2211
11-13-2003, 01:43 PM
Try to keep it above 30, bare minimum. The human eye can only distinguish FPS below that level, everything else looks relatively smooth. You should try to keep it above 50-60, though, so there's some leeway--more players in the map, especially bots, can decrease FPS.

jsera
11-13-2003, 07:15 PM
Thanks

idontlikegeorge
11-13-2003, 08:36 PM
With JA, I just try to make sure my framerate is not worse than the basic maps. For some reason, I get (relatively) bad framerates in Academy, even with special effects off - so I know if my map is getting really good framerate, I have some space to add details and such.

I think the r_speeds (remember that from DoD mapping) in this engine is r_showtris 1. I dunno what's reasonable amounts, since I'm still used to my HL mod mapping and wanting them at 1000 or less! :p

I'd say run r_speeds 1 on a Raven map of comparable setting and try to keep yours around the same.

Edited for having my foot in my mouth, metaphorically. Actually, r_showtris 1 draws the wireframe of the map... I guess.

Tyler_Durden
11-13-2003, 10:42 PM
I really wish the vaunted patch addressed this issue. Obviously it didn't happen but i'm hoping a 60MB patch that addresses this problem along with new mp maps, sp game fixes for nvidia owners and my long lost first person view shows up someday........

Lil Killa
11-14-2003, 02:54 AM
Originally posted by jsera
What is a reasonable FPS for a map?

I have a PII 800 w/ 512 mb and a 256mb nvidia card.




My friend has a comp about the same as yours and a map that normaly has 40 fps most people computers he gets something like 15 to 20 sometimes lower. When he played Bespin in JK2 he averaged at 10 fps with only 2 bots...

Jedi Luke
11-14-2003, 06:18 AM
I also like to keep my fps minimum rate at 30fps, usually i keep my fps between 50 - 60 fps depending on textures, scripts running, detail brushes etc.

ShadowBlight
11-15-2003, 12:10 AM
JA like all quake 3 based games, the frame rate is effected more by a proceser speed change then a upgrade in vidio card

I use a 1 gig athalon with 512 megs mem and a radeon 9000 with 64 meg memery

on small maps I can get 50-80 fps in some areas with out bots but add a few bots I drop to 15-30 fps

I sugest to improve fps turn off some of programs running in the background
if using a media player to play music wile you play, use winamp 2.x (it takes up less memery/processer cycles then the others)

Emon
11-15-2003, 07:41 PM
Framerate is a terrible way to measure performance, look at the triangles drawn with r_speeds 1. For singleplayer, it shouldn't be going much over 40K, for MP, not much over 25K.

yangyan
11-19-2003, 07:36 AM
is this related to mapping, or shouldn't an admin move this?

jsera
11-19-2003, 11:15 AM
Thank you all for answering my question.

Yes this post is about mapping. You should always make sure your map runs well on other machines and you need to use something to measure it.

Otherwise, you make the worlds largest or most complex room and people lag like crazy and dont play your map. No matter how beautiful it might be.

Thanks again.

shukrallah
11-19-2003, 12:34 PM
look at the triangles drawn with r_speeds 1. For singleplayer, it shouldn't be going much over 40K, for MP, not much over 25K.

ive never heard of "K" whats this mean? Is it in JA only? ill try out r_speeds 1 and look for myself.

JA like all quake 3 based games, the frame rate is effected more by a proceser speed change then a upgrade in vidio card

Also your RAM... i have a pIII 1.2 ghz, and 128 RAM, but i can only get about 30 FPS in normal maps... FFA_bespin varies between 15-25 FPS. in JA (I tried it out) I got about 13-26 FPS on yavin(If i remember correctly) but the yavin map was very detailed... very very nice in my opinion. (the first level) i walked half way through it, and got killed by a tree.

ewok mercenary
11-19-2003, 02:38 PM
K = Thousand

wedge2211
11-19-2003, 02:45 PM
"K" is just short for "thousand." "40K" is shorthand for "40,000." Comes from the metric prefix kilo, as in kilometer (1000 meters).

RAM does afffect things quite a bit. I noticed a very visible difference when I upgraded from 256 to 768 (also note: compiling is much nicer with more memory!) Video memory also does help.

Originally posted by Emon
Framerate is a terrible way to measure performance, look at the triangles drawn with r_speeds 1. For singleplayer, it shouldn't be going much over 40K, for MP, not much over 25K.
Well, as a counterpoint, frame rates can provide a very good indicator of the "general" performance of a map. You can't look at them and say, "this map gets 43 FPS while that one gets 47, so the second one has better performance," because the numbers aren't that accurate, fluctuate, and depend highly on system specs, but you CAN get a gauge of maps that have 15 FPS versus maps that have 90.

shukrallah
11-19-2003, 09:29 PM
oh... i thought it was a mapping term, lol. Anyways, i tried out the command on a small map i made, and i was getting like 145 or something.. there were a lot of numbers going down the console, so im guessing the first one was what im looking for. anyways, thats far away from bad. :)

Emon
11-19-2003, 11:38 PM
Originally posted by wedge2211
Well, as a counterpoint, frame rates can provide a very good indicator of the "general" performance of a map. You can't look at them and say, "this map gets 43 FPS while that one gets 47, so the second one has better performance," because the numbers aren't that accurate, fluctuate, and depend highly on system specs, but you CAN get a gauge of maps that have 15 FPS versus maps that have 90.

Measuring triangle counts or other stuff with r_speeds is the only way to tell how it'll run well on everyone's computer, because you keep the numbers within the accepted limits of current hardware. However you still need to check framerate for other stuff, like if you use blocksize that's too big, it can slow things like you wouldn't believe.

Lil Killa
11-20-2003, 12:21 AM
Originally posted by lukeskywalker1
oh... i thought it was a mapping term, lol. Anyways, i tried out the command on a small map i made, and i was getting like 145 or something.. there were a lot of numbers going down the console, so im guessing the first one was what im looking for. anyways, thats far away from bad. :)


it's one of the middle numbers..

Well that's what I have heard and what I use...

Emon. Wedge.. anyone?

idontlikegeorge
11-20-2003, 08:20 AM
I think it's the XXX/XXX TRIS part of the line. As to which number is what exactly, dunno.

Originally posted by Emon
Framerate is a terrible way to measure performance, look at the triangles drawn with r_speeds 1. For singleplayer, it shouldn't be going much over 40K, for MP, not much over 25K.

Would those ideals be for an all structural map - like a base map without entities/players, or should that be the max expected with entities/players/etc? Because after spawning in a group of reborns, both numbers before TRIS increased very noticeably.

Emon
11-20-2003, 07:17 PM
That's a good question. I would say before, but do a lot of performance testing during the beta stage, and if it becomes too much, tone it down.

idontlikegeorge
11-20-2003, 07:52 PM
Yeah that's what I was planning. I'm trying to make the basic map layout efficient, then testing it, then adding a level of detail to the environment, then testing it, etc.

I did spawn a buttload of reborns into my map, to see the performance with alot of NPCs... but, I was being an idiot, and didn't even turn on r_speeds... I did notice my framerate drop, but it was still playable, so yay! I got mail! Yaaay! :nut: