Hello.
I'm muddling over the following code, which compares an array/take
composition with the analogous imperative code. For medium-large values
of n, I'm seeing a fivefold degradation in performance, which blows up
to 30 times worse at n=50000000.
Any ideas on why this is or better ways to accomplish the same?
import std.range;
import std.date;
import std.random;
import std.array;
import std.stdio;
void main(){
for(int n = 500; n <= 500000000; n *= 10){
writeln(n);
auto r = rndGen();
auto tz = getUTCtime();
auto a = new int[n];
foreach(ref aa; a){
aa = r.front();
r.popFront();
}
auto tz2 = getUTCtime();
auto a2 = array(take(r,n));
auto tz3 = getUTCtime();
writeln("\tarr: ",tz2-tz);
writeln("\trange: ",tz3-tz2);
}
}
~
