Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

pwad speed and size limits. How big?

6 views
Skip to first unread message

Jason D Franklin

unread,
Apr 22, 1994, 1:52:52 AM4/22/94
to
WANTON is _TOO SLOW_ . It is also the largest PWAD out there
right now, at least as far as I know, so it makes sence that it would
also be the slowest. I noticed something interesting while playing
WANTON, however. The speed at the end of the level was much greater
than at the beginning. This was somewhat puzzling to me at first since
the level's size hadn't changed. The only change was in the number of
things and the type of things. Many of the monsters were now puddles
of blood and goo and bonus and weapon things had been picked up and
removed. Also, Michael Kelsey mailed to me a comparison of my level
and his CORTYR_B by size, # of sectors, # of things, # linedefs, ect.
and CORTYR_B though smaller in size has more linedefs and vertices.
The only truely outstanding difference was that WANTON had more
sectors and MANY more things. CORTYR_B is a very nifty level by the
way, good job Michael.
It seemed to me that the number of things definately played
a MAJOR role in the overall speed of the level, so I modified WANTON
to test this. WANTON has a large number of things in the middle
"arena" like area, especially the towers, so I altered those first.
Here are the results:

-486DX 33mhz

WANTON (unchanged)

Beginning of level: 6-7 dots.
Running in "arena" area: 4-6 dots, with pulses up to 9 when
a monster was "awakened".
At door looking into skull area:5-6 dots.


WANTON (monsters in towers now stim packs)

Beginning: 4 dots.
Running: 3-4 dots.
At door: 4-5 dots.


WANTON (no monsters in towers)

Beginning: 3-4 dots
Running: 3 dots.
At door: 4-5 dots.

It seems that the number of things is an important factor when
determining speed of a level especially when they are in the field of
view of the player, and even if they are behind a wall. Also, monsters
seem to slow things down a bit more than other things. In order to get
a reference point for this data here are the results for Mt. Eribus.

Mt. Eribus

Beggining: 2-3 dots
Running around: Varies greatly 2-4 dots pulses to 7.


To speed up WANTON I split it into two PWADs with the same map
but each containing only around 1/2 of the things. The exit for E1M1
is the beginning to the third "quadrant". You'll have to look at the
level to see what I mean. The results of thse levels are:

WANTON12 (E1M1)

Beginning: 2-3 dots
Running around: 2-3 dots pulses to 5


WANTON12 (E1M2)

Beginning: 4-5 dots
Running around: 3-5 dots pulses to 7


So, if you can play Mt. Eribus, you should be able to play
WANTON12.

Now, about size- DO NOT attempt to make a level the size of
WANTON with DEU 5.0. It will crash repeatedly. When I was about 3/4
through with WANTON DEU 5.0 started crashing after simply inserting a
few linedefs of vertices. It has to do with the way DEU reallocates
after the creation of every linedef, sector, or vertice. The only
reason I was able to complete WANTON is that I happen to have a
bitchin' programmer friend named Bryan Craig Smith. He patched DEU's
realloction scheme and I had no problems at all after that. He has
talked with the authors of DEU and the next release should contain the
patches. So until then, COURTYR_B is about as large as you want to
get. But that is plenty big.

Does anyone know anything else about these issues?

jay...@cs.tamu.edu


Leo Lim

unread,
Apr 22, 1994, 5:53:32 PM4/22/94
to
Interesting observation.

I am working to speed up wanton.wad myself.
Tho, my fix was not reducing number of things, but
I deleted several vertices/linedefs/sectors that you won't pass anyway.
look at the whole level closely, you can find
it has 4 HUGE triangular sectors covering 4 sections of level and 1
GIGANTIC rectanglar sector covering the whole level.

also, I have changed some of the linedefs flags, such as block sounds etc,
since alot of them are redundunt/un-necessary, this changes didnot affect the
original intention.
also some sectors floor/ceiling height of the arena area.


I haven't finished fixing all yet but, so far it has sped up significantly
over the original.
too bad i don't have DotsMark(TM) handy, sorry. :)
i'll post more after i finish fixing it. (hey, I want to keep this wad)

To whoever build this wad: Nice Idea, but need some tweaking.

===Martin

0 new messages