Re: Request for timing information
jon klein <[email protected]>
| Newsgroups | gmane.comp.breve |
|---|---|
| Message-ID | <[email protected]> |
On Mar 4, 2004, at 12:47 PM, Eric DeWitt wrote: > I am writing genetic code that must be run in every agent (potentially > a large number) and requires some nested looping over what could be a > large number of elements. I need to get some accurate information on > timing of some operations in breve/steve and I think that they'd be > useful for everybody (and might end up in the documentation > eventually). What I'd like is information on the cost of the > operations mentioned below in both some absolute and relative terms. > (For example, is the cost of the indexed retrieval from a list O(N), > with what constant and how does this constant compare to that of > method call overhead?) Some of this is, I am sure, constant time and > insignifigant--I just would like some relative idea of the costs. I > assume a pretty simple test could be configured to perform a large > number of each operation that doesn't have a clear answer and the > profiling information extracted and averaged. This is a very complicated question. I think everything you mentioned has constant time complexity, but I have no figures for the relative time that each of these operations takes. I would suggest you look at the source code, or try running breve with a profiler. Apple's "Shark" profiler (part of the CHUD tools) is really excellent for this type of thing. - jon klein _______________________________________________ breve mailing list [email protected] http://www.spiderland.org/mailman/listinfo/breve