No, that's exactly what's happening. I was just confused by the active verb which struck me as you might have done something to compile the code.
On Sun, Jul 6, 2014 at 6:10 PM, Michael Louwrens < [email protected]> wrote: > Previous message was rubbish sorry. > > I wanted to test how well successive calls would run. > > 1st call was 100ms 2nd call was the far lower one. I assumed that it was > because the JIT had optimised the function already? Or was I sorely > mistaken. > > > On Monday, 7 July 2014 03:02:00 UTC+2, Stefan Karpinski wrote: > >> This is the bit I was confused about: >> >> On Sat, Jul 5, 2014 at 12:19 PM, Michael Louwrens <michael.w...@ >> outlook.com> wrote: >> >> >>> Interestingly, Julia gets 100ms if I simply use @time testFunction(), on >>> par with c++. >>> >>> Using >>> testFunction() >>> @time testFunction() >>> >>> To compile speeds it up to 57ms, the fastest of all the languages tested! >>> >> >> "if I simply use @time testFunction()" implies to me that you were doing >> something else before. I was wondering what that something else was. I'm >> also confused about "To compile speeds it up to 57ms" – what does that >> mean? How did you compile it? Julia always runs compiled code... >> >
