OK I have compiled my map for CSS fully and when walking around just by myself the map is smooth and fps are good to great (45+ to 150+).

BUT, when I added several bots to the mix to test for playability my fps saw a severe plummet in "higher resource" areas (fps below 20 in some spots mostly 35ish).

I have read up on optimizing and so fourth but my question really is: Does my map really need that much more optimizing when its good with one player? OR is it just laggy because my computer has to control bots AND draw the map?

This map is intended to be used as a multiplayer CSS map with (i was hoping) up to 20+ people. I'm running a 3Ghz PIII 512Ram ATI radeon x700.

Thanks for your time and advice.

BENDER

45 fps as one man doesn't really say "great FPS" to me. It'd be nice if
you accessed the console and showed us shots of the most detailed areas
with the showbudget command.

I think you should perhaps post your map as "99% complete" in the maps
forum if you are looking for optimisation. Without specific errors, it
is too general and large an issue to cover in one thread.

What he said, but worded less coherently and with more rambling.

---

o

With an engine as powerful as this, there will always be a wide range of frame rates.

The best thing you can do is post it as Gwil suggests, and allow a wide range of machines decide if your map is playable.

I consider anything over 30 acceptable "IF" there are no blatant errors in your map forcing low frames.. (IE poor craftsmanship)

---

The best things in life, aren't things.

FPS alone don't say anything. It only helps you to compare a map with others. If 45-150 FPS is what you get on a default map (say de_dust) then it's fine for your map too. If you get less FPS with the same settings, you should consider a little more optimization.

FPS are usually measured without players for map testing. But it's generally a good idea to build your map in a way that not all players see each other constantly. Both to reduce internet-lag (ping) and FPS.

Personally, I do not think that the CPU load of Bot AI affects your FPS much. Maybe 3 or 5% but not a drop from 45 to 20.

Ok first, thanks for the info, but I guess to scale my question back a bit from my own personal map, here's a better question I suppose:

Regardless of what my map looks like (complexity,detail,layout,etc.) will AI bots (10 or more) slow down performance/fps THE SAME as real people multi-players?

I ask this question because I have to disagree politely w/reaper's quote about bots affecting performance (at least on my system). All alone in my map the animation is 40+ish fps (only in the worst spots) but with bots fps PLUNGE to 5 to 20 and I think my screen shots prove it.

<DIV class=quote>
<DIV class=quotetitle>? quoting reaper47</DIV>
<div class="quotetext">Personally, I do not think that the CPU load of Bot AI affects your FPS much. Maybe 3 or 5% but not a drop from 45 to 20.</DIV></DIV>



NO BOTS (47fps)



WITH BOTS ( 3 fps would vary into the teens)



NO BOTS (47fps)

http://i21.photobucket.com/albums/b258/tjbender/junkyard_0070005D.jpg

WITH BOTS (11fps with some varyance)

I just want to know if the choppyness i'm experiencing with the bots will happen with real people multi-players, because if so, i need to revamp quite a bit (hence the development meter at 80%).

also of interest, I noticed performance/fps INCREASE as bots died...Hmmm.

You guys are the best, thanks again for your QUALITY help!

~BENDER

::EDIT:: Reap, I will try one of the regular maps w/some bots and see what i get. i havent played them w/ bots in a long time so i dont know. i'll try militia cuz i know how dense that is w/models and such.i'll report back soon.

</div>

The only way bots effect FPS is if your machine is unable to handle the workload of rendering them. Each machine will be effected differently by the amount of models rendering in each frame.

My old 9800 would prolly not even notice 10 bots. My 9000 will grunt a bit.

This is the problem as I see it.

I'm running a 3Ghz PIII 512Ram ATI radeon x700.

You are running a very old machine.. Its that simple.

Find someone with a bigger machine to give you a better sense of how the numbers will be portrayed with a more modern machine... Then worry.

/2cents

(I am making this yellow only because I am nearly certain of my reply. Normally I would forego the yellow.)

---

The best things in life, aren't things.

I just had this highly experimental idea of how to test whether bots slow down gameplay more than people. Bots have to calculate AI (the only thing that's different from player-controlled characters) and they have to do even if you don't look at them (I guess). Walk to a wall facing outside the level and turn your back to all the bot action. Make sure you're looking at a wall with nothing even behind it. That way you get FPS nearly unaffected by the models rendering.

Now see if you still get a major performance decrease. It should only be the AI in the background then, throwing their grenades and thinking of fitting voice commands. If FPS suddenly increase again like in a non-Bot match then it's just the model rendering and not the bot AI.

From the screenshot I see places that could be optimized quite a bit I guess. But for that I'd open a new thread in the maps forum.

ok, I get the fact that my machine may be too slow to handle 10+ bots and my map at the same time......BUT....does the fact that the map runs slow on my comp w/bots ALSO mean that it will run slow on a good server w/real people?

It's not slow with just me running around.

Do real people multi-players slow a map up as much as bots do?

Thanks for everyone's patience.

::EDIT:: Reap I just read your post. OK I totally get what you're saying about your "experiment" and its true...when looking at a wall away from action fps are like 150+, but they are just as good without the bots too though. I just feel that the bots are slowing the game up SO MUCH more than the inefficiencies of the map, MY PROBLEM/QUESTION REALLY IS will these bot slow-downs translate into real people multi-player slowdowns. I don't know. thats why i'm here. whew! Thanks again. BTW i was incorrect about my processor its a P4 w/x700pro, not that it prolly matters, my comp isnt that archaic.

My guessed answer is that you'll get the same FPS with 10 online players. The only way to find it out is... testing it with 10 online players. And probably asking them if they have similar performance troubles.

I looked at some of the screens from the map profile. This map looks very open with little brush architecture to block the view. That's always bad for FPS. Use more circular layouts instead of a "big box" with stuff inside. This keeps at least half of the player models from being drawn most of the time.

Optimization can be a hard task if you don't plan it from the beginning. Read this tutorial carefully. It explains everything you need to know about map performance:

[url]http://www.student.kun.nl/rvanhoorn/Optimization.htm [/url]