Rediculous Compile Time

I've been compiling my competition map since 2am thursday morning and the thing is still going. It took 30 hours to get to 9... and its been stuck trying to get to ten for the last 36 hours. Does anyone have a clue, because I have to send this thing on monday or thats it, a months worth of work for the comp down the drain.

PLEASE HELP ME IF YOU CAN!

Omg... I just came on to post the exact same thing... is this what your compile looks like?

materialPath: c:\program files\valve\steam\steamapps\ttevil\counter-strike source\cstrike\materials
Loading C:\Program Files\Valve\Steam\SteamApps\ttevil\sourcesdk_content\cstrike\mapsrc\cs_alcatraz503.vmf
Detected multiplayer map, building linux dedicated server data
fixing up env_cubemap materials on brush sides...
0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . . 10 0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . . 10 Processing areas... done (0)
Building Faces... done (1)
Chop Details... done (0)
Find Visible Detail Sides...
Merged 358 detail faces... done (0)
Merging details... done (0)
FixTjuncs...
PruneNodes...
WriteBSP...
done (6)
writing C:\Program Files\Valve\Steam\SteamApps\ttevil\sourcesdk_content\cstrike\mapsrc\cs_alcatraz503.prt... done (0)
Creating default cubemaps for env_cubemap using skybox sky_day01_09...
Run buildcubemaps in the engine to get the correct cube maps.
Finding displacement neighbors...
Finding lightmap sample positions...
Displacement Alpha : 0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . . 10
Building Physics collision data...
done (3) (1093547 bytes)
Building Physics collision data...
done (2) (1171680 bytes)
Placing detail props : 0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . . 10
Writing C:\Program Files\Valve\Steam\SteamApps\ttevil\sourcesdk_content\cstrike\mapsrc\cs_alcatraz503.bsp
34 seconds elapsed

1 threads
reading c:\program files\valve\steam\steamapps\ttevil\sourcesdk_content\cstrike\mapsrc\cs_alcatraz503.bsp
reading c:\program files\valve\steam\steamapps\ttevil\sourcesdk_content\cstrike\mapsrc\cs_alcatraz503.prt
1487 portalclusters
3661 numportals
0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . . 10 0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . .

Pretty much exactly the same, except I've got less portalclusters and more numportals. It's been sitting there, using 99% system resources since early thursday morning, trying to build that last segment. I can't tell if its frozen or what.

I got an error the first time I tried to do a full compile, something about too many portals, which I fixed. I did a quick compile before hand to check for anymore errors and stuff but it was clean. I put on the full compile and its still going. No errors or anything.

Mine was working fine until I made the skybox bigger and added a displacement island.

I am going to try and find out why... if I ever do find out, i'll tell you straight away.

I've been compiling my competition map since 2am thursday
morning and the thing is still going. It took 30 hours to get to 9...
and its been stuck trying to get to ten for the last 36 hours. Does
anyone have a clue, because I have to send this thing on monday or
thats it, a months worth of work for the comp down the drain.

PLEASE HELP ME IF YOU CAN!

You need to learn to use GLView. It's your VIS that's killing your compile time. I bet your leaves look like hell.

Wait, there were some errors I missed, could these be the problem?

2 threads
reading g:\steam\steamapps\agent_smith_bob\sourcesdk\dm_hydro.bsp
reading g:\steam\steamapps\agent_smith_bob\sourcesdk\dm_hydro.prt
1879 portalclusters
6534 numportals
Leaf (portal 1835) with too many portals. Use vbsp -glview to compile, then glview -portal -portalhighlight X to view the problem.

2 threads
reading g:\steam\steamapps\agent_smith_bob\sourcesdk\dm_hydro.bsp
reading g:\steam\steamapps\agent_smith_bob\sourcesdk\dm_hydro.prt
1809 portalclusters
6245 numportals
0 . . . 1 . . . 2 . . . 3 . . . 4 . . . 5 . . . 6 . . . 7 . . . 8 . . . 9 . . . 10 WARNING: Cluster portals saw into cluster
WARNING: Cluster portals saw into cluster
WARNING: Cluster portals saw into cluster
Optimized: 148516 visible clusters (0.00%)
Total clusters visible: 2762722
Average clusters visible: 1527
Building PAS...
Average clusters audible: 1800
visdatasize:819952 compressed from 839376
writing g:\steam\steamapps\agent_smith_bob\sourcesdk\dm_hydro.bsp
18 seconds elapsed

1. I don't think them errors are the cause.

2. There is nothing wrong with my leaves... you need to learn ask people whether they use it or not instead of telling them they don't

It's taking me like 2 minutes at the most to get to 1...2...3...4...5...6...7...8..9... but then even if I leave it for 2-3 hours it doesn't go to 10. This can't be a problem with leaves can it if it only takes 2 minutes to get to 9...?

Holy crap your map must be huge, but there is no reason why it should take over 24 hours to compile. I'd cancel it right now and do some quick simplification of complex areas and then retry. Cut unnecessary detail where you can.

I can tell you that the WARNING:cluster portals saw into cluster is either meaningless or not important, as my map has that and it finishes vis in about 15 minutes.

I would look into the leaf w/ too many portals error, as that could be messing everything up.

Sorry I can't be much more of a help.

Even if u have your map in a sky box and u have leaks that will leak inside the sky box but not outside of it, then Vis has to calculate everything outside of your map but inside the skybox. U have to delete a wall of the skybox aka 'Brush" and then compile the map, if u have leaks then load the MAPNAME.lin file(it get's generated only when u have leaks) that will be in the hammer menu Map -> Load Point file and it will show u where are the leaks. TIP every displacement that i applied to a a solid LEAKED then i had to enclose only parts of my map in skybox.

Can someone check out my .vmf file and see if they can fix any problems for me... it is not very optimized though as I kept having problems of portals refusing to open.

If you can add livett_1@hotmail.com to MSN and I will send it to you

<DIV class=quote>
<DIV class=quotetitle>? quoting alingrecea</DIV>
<DIV class=quotetext>Even if u have your map in a sky box and u have leaks that will leak inside the sky box but not outside of it, then Vis has to calculate everything outside of your map but inside the skybox. U have to delete a wall of the skybox aka 'Brush" and then compile the map, if u have leaks then load the MAPNAME.lin file(it get's generated only when u have leaks) that will be in the hammer menu Map -> Load Point file and it will show u where are the leaks. TIP every displacement that i applied to a a solid LEAKED then i had to enclose only parts of my map in skybox.</DIV></DIV>
I'll try that out. Good advice :biggrin:

Wait a minute... how am going to try that if my map won't compile?

The new vis is a bitch. Seriously, if you have anything complex vis-wise, you have to optimize like crazy or else the vis times just skyrocket. I'm just trying to find out whether the vis algorithm actually has to take into account all the previous stuff as it goes along, or whether it just does the easy stuff first and the hard stuff last. Hopefully its the latter since then you could just blockify all your complex stuff with hints and it would speed up.

I found that when I deleted the island and made my skybox smaller my building was on the VIS time was a hell of a lot faster which makes me think I need to vis block my building from the rest of the map.

I tried doing this but I couldn't get it to work, also I can't get my area portals to work properly. Any chance I can send you my .vmf and you optimize. I try to learn to optimize whilst I am waiting for a reply. Hopefully I can learn to do it myself.

hmm, just need to say that im pretty much a mapping noob...
But when im compiling the my map, just to try it out, it gets stuck on one of thoose meters (1...2....3...) At first it goes okay, but it gets slower and slower until it stops.

So, i got win xp, i checked at the ctrl+alt+del window (dunno the english name) and i found a process, "vvis". And it took like 99% of my cpu, I din't know what it was, but i dont want no processes taking 99%. so i shut it down. and suddenly hl started. becuse that compile stuff got canceled so the rest could go on. When i tested the map, nothing was wrong or anything and it played well.

what im trying to say, it that um, I can play my map without letting vvis work for more than a minute or so. So you might try that as a last resort. Just to get you map playable until monday. Just quit the process, and se what happens.

But! remember that im a mapping noob, and i dont know what vvis does or anything! So be cautios if you try that....

Of course it takes up so much CPU. If it only used up like 5% then it would take so much longer.

It is best to let VIS run through, or you will get serious problems with your map.

the thing is that i dont, no problems at all... But im sure it has a prurpose, so i guess its best to wait like all day to compile it..... :/

I've done some fixing up so hopefully it'll run a bit better. I did
notice however when looking at the vis leafs that there seemed to be
some outside the actual level, going of into the distance. Does
everyone else have them or is something f**ked in mine?

Can't you try running Vis on fast?

i've compiled maps and the difference between fast and normal i've seen is minimal....

The difference will only be minimal in certain maps - in most maps,
particularly those that aren't just large open areas, the difference
will be quite significant.

Everythings under control now. The compile is still gonna take about 24-28 hours, which is a bloody long time, but significantly shorter than the previous. I did some optimization, added some hints and found out where all the dodgy leafs were and fixed 'em up and it seems to have worked.

I did however reduce the height of the sky in the open areas to minimize the leafs present (imagine making a big sky box around the level, then in sections of blank space place another sky box but have it inverted, so the inside leads outside the level, reducing the leaves for that section), and as such the height of the sky for weapons isn't as big as before. The weapon stopping roof of the level is about in the middle of the cranes arm (see DM_Hydro in my profile for screens, should have up some new ones when the compiles done. Still a decent height, but not as high as before. Rockets also blow up if you hit it and grenades bounce off.

Oh, and fast compile takes about 5-10 seconds on my computer for VIS.

You don't necessarily need to lower the height of the sky - just make a
hint brush filling the skybox above the heighest point of your
brushwork and this will mean it will only use one or two leafs.

Thanks Reno, never been quite sure how to use hints. Since my brush work goes to the top of the sky as it is, I just layered two large hint brushes at the top of the sky box, going down to but not quite touching the more detailed areas. It's sped up the compile even more.

In addition, make sure that you have tied most of your brushes to
func_detail. It has the greatest potential for speeding up
compile time. I've spent over a month making a fairly large map,
and it still only takes fifteen minutes for it to compile.

I'd say if everything is optimized the way it should, no map should
take longer than half an hour to compile. If the size of the map
gets larger, then the map should be split into multiple section to have
a reasonable compile time and performance.

<DIV class=quote>
<DIV class=quotetitle>? quoting Livett</DIV>
<DIV class=quotetext>
<DIV class=quote>
<DIV class=quotetitle>? quoting alingrecea</DIV>
<DIV class=quotetext>Even if u have your map in a sky box and u have leaks that will leak inside the sky box but not outside of it, then Vis has to calculate everything outside of your map but inside the skybox. U have to delete a wall of the skybox aka 'Brush" and then compile the map, if u have leaks then load the MAPNAME.lin file(it get's generated only when u have leaks) that will be in the hammer menu Map -> Load Point file and it will show u where are the leaks. TIP every displacement that i applied to a a solid LEAKED then i had to enclose only parts of my map in skybox.</DIV></DIV>
I'll try that out. Good advice :biggrin:

Wait a minute... how am going to try that if my map won't compile?

</DIV></DIV>
well you can compile without vis and rad and it will compile fast and it will show u errors. other than that i have no other good advice. And i think when u have a leak vis automaticaly stops calculating (when the leak goes out in void ) when there is no skybox.

<DIV class=quote>
<DIV class=quotetitle>? quoting satchmo</DIV>
<DIV class=quotetext>In addition, make sure that you have tied most of your brushes to func_detail. It has the greatest potential for speeding up compile time. I've spent over a month making a fairly large map, and it still only takes fifteen minutes for it to compile.

I'd say if everything is optimized the way it should, no map should take longer than half an hour to compile. If the size of the map gets larger, then the map should be split into multiple section to have a reasonable compile time and performance.
</DIV></DIV>
yes but they need to know the func_details dont block vis so it can be tricky.