Zugg Software :: View topic - Which is quicker: one trigger or many triggers?

Wanderer Joined: 30 Aug 2001 Posts: 53 Location: Finland

Will zmud work more quickly if you use only one trigger:
#tr {car|cat|cool} {#co bold,blue}
or if you put them in different triggers like:
#tr {car} {#co bold,blue}
#tr {cat} {#co bold,blue}
#tr {cool} {#co bold,blue}

?

GURU Joined: 10 Oct 2000 Posts: 4379 Location: USA

The second form should be faster but the difference shouldn't be noticeable. Also, in the latests versions of zMUD, the first form had some optimizations made and thus the difference between the two is even less.

Kjata

SubAdmin Joined: 18 Nov 2001 Posts: 5182

From what Zugg has stated of the optimizations I expect the first form should be faster as of the next public release (or current beta if you wan't to risk your settings). In the current public version, the second form is slightly faster. Of course I can not not be definite on that without violating the liscense agreement.

GURU Joined: 10 Oct 2000 Posts: 4379 Location: USA

The first one is faster than it previously was in the current beta because of optimization, but it still is not as fast as using multiple text-only triggers. The reason for this is that any trigger that contains any wildcard or pattern expression (like the {list} syntax) gets sent to a different, slower parser that text-only triggers. More specifically, zMUD first tries to discard a line using any text that the pattern of the trigger contains. If the line is still not discarded and the pattern contains any wildcards or pattern expressions, it gets sent to the other parser. With the second form, none of the triggers would ever reach the other parser while the first form would get sent straight to it. However, it should also be noted (and I can't back this up as I didn't code zMUD) that, in the current beta, as the size of the list grows (meaning more and more separate triggers in the second form), the first form would be faster since zMUD doesn't have to go agaisnt each trigger checking for a match, it just sends the whole list to the second parser which will create a hash table of all the items in the list and use that to find a match. But, again, the difference in speed shouldn't be noticeable at all.

Kjata