Take a look at the performance tips in the manual: http://docs.julialang.org/en/release-0.5/manual/performance-tips/ <http://docs.julialang.org/en/release-0.5/manual/performance-tips/>
One of the main things is to avoid accessing global variables, and everything declared at the top-level in the REPL is global, so you generally want to put any performance-sensitive code (and variables accessed in that code) inside a function. -s > On Sep 24, 2016, at 12:04 PM, Weicheng Zhu <ming...@gmail.com> wrote: > > Hi there, > Could anybody tell me what's wrong with my for loop code below and how to > improve it? > I was surprised to see it performs worse than the vectorized operation. > Thank you! > > ## Calculate the maximum number in vector x that is less or equal to 0.5 > ## Method 1 > x=rand(10000) > @time maximum(x[x.<=0.5]) > # 0.000205 seconds (23 allocations: 45.141 KB) > ## Method 2 > u= x[1] > @time @inbounds for i in 2:length(x) > if u < x[i] && x[i] <= 0.5 > u = x[i] > end > end > # 0.001233 seconds (42.40 k allocations: 818.766 KB) > > > Best, > Weicheng > >
