On Dec 7, 2012, at 9:42 PM, Andy Fingerhut wrote: > > > When you say "we can run multiple instances of the test on the same machine", > do you mean that, for example, on an 8 core machine you run 8 different JVMs > in parallel, each doing a single-threaded 'map' in your Clojure code and not > a 'pmap'? > > And what kinds of speedups does that achieve? > > The results could help indicate whether the problem you are seeing is due to > the hardware/OS, or something about multiple threads within a single JVM.
Here are a couple more results, courtesy my colleague Josiah Erikson (who may chime in here at some point too): On a 4 core machine, each test within a single "lein run" call: 2:31 for (doall (pmap burn (range 8))) 1:29 for (doall (map burn (range 8))) 1:48 for (doall (pmapall burn (range 8))) [see my reply to Jim, immediately following this one, for pmapall] Now with multiple "lein run" calls launched concurrently: 29 seconds to do 4 concurrent copies of (doall (map burn (range 2))) 33 seconds to do 8 concurrent copies of (burn 2) [the argument to burn doesn't matter, but that's what Josiah ran] 38 seconds to do 8 concurrent copies of (burn (rand-int 8)) [again, the argument shouldn't matter]. 48 seconds to do 2 concurrent copies of (doall (map burn (range 4))) 1:07 to do 2 concurrent copies of (doall (pmap burn (range 4))) What do we learn from this, aside from the fact that things are way wacky? Thanks for the help!! -Lee -- 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
