Sounds like you might have been running into the absence of multimethod caching for the default case (http://dev.clojure.org/jira/browse/CLJ-1429), which has been fixed in 1.7.
Just on a side note, that repo does not change the default lein :jvm-opts, which are bad for benchmarking. I recommend at least: :jvm-opts ^:replace ["-server"] Criterium now yells at you if you haven't done this. On Monday, April 27, 2015 at 9:43:11 AM UTC-5, Phillip Lord wrote: > > > I think that the answer is, it depends, and, there might be some > surprises in store. > > In my own use, I found multimethods collapse in performance as a result > of changes to the interface hierarchy in the library I was using (my > tests cases went from 45s to over 4mins). > > I circumvented this by fiddling with the dispatch function, so that it > returned ONLY classes that exactly matched one of the ones in the > multi-method, and memoizing the look up of these classes. In effect, > this closes down the multi-method, so it can no longer be freely > extended. This dropped the time of my test cases dramatically. > > How widely applicable these results are, I do not know, as I never got > down to the mechanistics underneath. Still, my conclusion is, you might > want to test things out first! > > My full set of experiments are here... > > http://www.russet.org.uk/blog/3007 > <http://www.google.com/url?q=http%3A%2F%2Fwww.russet.org.uk%2Fblog%2F3007&sa=D&sntz=1&usg=AFQjCNFnegld_FgM0oFf4JgU9a1UvAAFaQ> > > > > Timur <timur...@gmail.com <javascript:>> writes: > > > Hi everyone, > > > > There are situations where I want to dispatch functions using based on > > their certain properties. I can also use case statements instead but it > > looks more coupled and more change is required if I want to add new > types. > > > > What I want to ask is if I need to avoid using multi-methods for > > performance reasons? I read somewhere that they are not really fast but > the > > posts were old and the performance might have been improved in between. > > Should I favor case and cond branches instead of defmulti when I need > > performance? > > > > Thanks for your help!!! > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
