Compile Times Reaching Ludicrous Speed^-1

Anpheus · November 2004

Most recent RAD compile:

20690.09 seconds elapsed [5h 44m 50s]

I have done everything I percieved possible to reduce this time. This includes making a big box around the map to check for lighting but non BSP breaking lighting leaks, this includes turning half my map into a collection of func_walls for the purpose of improving VIS. This includes creating 1 to 2 unit wide spacer func_walls to surround large brushes to prevent face-splitting. This includes removal AND addition of lights, and numerous other things.

Despite all this, compile time has been growing steadily. I request expertise assistance for this, because <b>five hours</b> for a single (although exceptionally large) room is excessive. Compile time has been as low as three minutes in the past.

Pithlit · November 2004

first remove the "big box" around your level
fixing leaks with such a thing is evil and boosts your compile times enourmusly, since now the complete outside of your map has to be lighted aswell-

Anpheus · November 2004

I've long since removed the big box. It was a test because I had been told that it may reduce compile time. It didn't, so I removed it. The leaks I am talking about are light leaks, which are caused when two brushes intersect in a line or a point, which sometimes allows light to go through the point- or line-intersection. I was told that creating a big box may help this. It didn't.

There are <b>ZERO</b> leaks in my map. I will not reinforce this any more.

Now, expert advice, anyone?

BulletHead · November 2004

erm.. let me see the .rmf or at least a few screenshots... I'd like to know what in the HELL can cause a compile time like that for one room 0o'

How many OPAQUE entities?
How many lights, light_spots, and the like?
How many brushes?
HOW BIG?

Show us the thing yer trying to compile and we can help more <img src='http://www.unknownworlds.com/forums/html//emoticons/smile-fix.gif' border='0' style='vertical-align:middle' alt='smile-fix.gif' />

Then again... my current map in progress is like... 6 hours + to compile on FAST vis with NO opaque and NO extra... and I think it's around 12 hours with FULL, OPAQUE, and EXTRA 0o'

Anpheus · November 2004

Fairly large, it's a pyramidal room with a minimum of brushwork that has five 'depressions' led down into by staircase. It's meant for the top of a large building. It has a main center depression ringed by two 20-unit tall stairs made from a 360 arch. Surrounding this is are some worldbrushes to make the area a square. Surrounding the worldbrush is a one unit wide func_illusionary. Surrounding this is eight more brushes that combined create a near-square with the four corners missing. In these corners is a small descending area with another staircase, this time composed of four stairs that are all walkable. These four smaller staircase areas are slightly lower than the level of the larger platforms, and they are surrounded by sixteen brushes in various func_wall entities, which then completes a full square around the outside of the main ring. Then, the square descends a little bit to reveal a texlight from a failure of a texture. The texlight rings the entire area, and is part of the outter shell of the pyramid. The pyramid also contains four columns, one in each corner, which are composed of two square mitered worldbrushes and two angular mitered worldbrushes, these columns are seperated from floor and ceiling by a one unit tall func_illusionary. On these rests a balcony, which due to the pyramidal shape, also hangs from the shell of the pyramid. Also on each column is a stretched pentagonal shape which extends towards the center for a square center platform. Between the pentagonal shape and the balcony is another one unit wide func_illusionary, so really the pentagon shape only connects with the square wolrdbrush in the center. And between each column there are two 'hanging' platforms, supported by func_walls, with func_wall stairs below the platforms.

Additionally, above this square platform is a smaller room, in which there is an inverted pyramid at the top that emits light into this small room. The small room is opened into by a 16-vertex circle, composed of two worldbrushes and two func_walls.

Additionally, for every connection between the pyramidal shell of the map and a world brush within, there is a func_illusionary between the shell world brush and the internal world brush.

I think I got everything. I'll show pictures when this current compile is done, which can be any time between now and tomorrow.

Edit: Oh yeah forgot the opaque and light information. There are no opaque entities, and there are probably 12 lights, and a bit of texture lights. The lights are definately not the problem because I have had them for a long time, and it wasn't them that caused the extreme rise in compile time.

Edit: The problem, I have found it. It was the texture lights that I used. Specifically, I had shrunken the texture so, presumably, RAD would chop it into finer patches.

carioca · November 2004

<img src='http://www.unknownworlds.com/forums/html//emoticons/biggrin-fix.gif' border='0' style='vertical-align:middle' alt='biggrin-fix.gif' /> what are your computer and try reduce the light power, less light less compiling, try down the gama value in the env_gama to make the darknessbright more.

MADMEE · November 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (carioca @ Nov 8 2004, 02:13 AM)</td></tr><tr><td id='QUOTE'><img src='http://www.unknownworlds.com/forums/html//emoticons/biggrin-fix.gif' border='0' style='vertical-align:middle' alt='biggrin-fix.gif' /> what are your computer and try reduce the light power, less light less compiling, try down the gama value in the env_gama to make the darknessbright more.</td></tr></table><div class='postcolor'>
er.. no. the gamma ramp is not taken into affect in the compile else you wouldnt beable to change it on-the-fly after the map is compiled. it is an ajustment of your hardware.

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> </td></tr><tr><td id='QUOTE'> this includes turning half my map into a collection of func_walls for the purpose of improving VIS. This includes creating 1 to 2 unit wide spacer func_walls to surround large brushes to prevent face-splitting.</td></tr></table><div class='postcolor'>

i wouldnt do any of that either. seeing as entites dont block light all those func_walls arent helping they are just making it calculate the func_wall and the brush behind it.

and some screenies from vhe would be more helpfull then that rather confusing discription of non-square square things.

also did the compile complete or did u stop it? as when i was compiling my map it was taking 5 hours to realise it was out of light memory. however once i increased the amount of memory available it only took 3 hours.

Anpheus · November 2004

I found the solution, and edited my most recent post to declare this. I did that... two days ago?

Compile Times Reaching Ludicrous Speed^-1

Comments