What would put less strain on a server, 25-30 tiny maps or one huge map? If there isn't a difference, I guess it really doesn't matter. But I am mapping from Pokemon FR/LG and was wondering if putting all of Kanto in one map was a good idea, or should I include z map transistions from route to town?
Thanks for any comments.
ID:166632
 Jul 11 2006, 8:46 pm
|
|
if you play alone it doesnt matter how big your map is.
If there are going to be a lot of players on your map(unlikley from what I've seen from MOST other games)it would reduce lag by using muliple maps, do a big one, test it with a few friends then decide. | |
The world doesn't change maxx, maxy and maxz at startup with multiple map files compiled in any more than it does with a single one.
All the map data is compiled in at compile-time exactly how it will be in the running world no matter the number of maps you use; it's all done at compile-time, not at startup. Therefor, it does not change compile time or startup time if you work with one bigger map or with multiple smaller ones. To test this assumption, I just made a test world that had one 1000x1000x10 map, then I compiled and timed it, then I changed it to one 1000x1000x1 map included and one 10x10x9 map included. The time taken to compile and the time taken to start the world were the same in both cases. They both took nearly 7 seconds. In the end, the only difference whatsoever that you should have between using one map file or multiple map files is purely organizational. If it's easier for you to organise multiple map files than to navigate along the z-axis of one file then go for it, but that is the only benefit you will get from it. | |
That's odd, as I noted a distinct difference between the loading times at runtime and at compile-time when I tried it.
Must be my computer. Sorry! | |
That is true, there could be more factors involved. For example, my test example was just a vast field of default /turf, so maybe the startup time would have been different if I had many more types. I don't know, but either way it just seems insignificant enough that it doesn't matter unless you are working with millions of tiles.
When I changed world maxx/y/z at runtime, it took approximately 1.3 seconds per million tiles I would add or remove, but to compile or startup-load a map of the same size would take approximately two-thirds that amount of time. Therefor, at least for my computer, any time seems insignificant unless you have a truely vast map. And my computer is only middle of the road. | |
On the other hand, when you have a game which you want to compile a lot, having multiple maps proves to be more productive as it will decrease compile time because it doesn't need to bother about wasted space. However, on runtime there will be a speed decrease because it needs to adjust world.maxx, world.maxy and world.maxz while looping through the maps to add.
It's not that much of a difference in speed, but enough to get noticed. Therefor, it's basically a matter of preference, really. I tend to use multiple maps so that compile time is decreased enough so I can still work on things without having to wait eight seconds per compile.