Performance of tables slower than built in?

[Alex]

On Thursday, 23 May 2019 at 19:17:40 UTC, Ola Fosheim Grøstad 
wrote:
> On Wednesday, 22 May 2019 at 00:22:09 UTC, JS wrote:
>> 	xxx = 0;
>> 	sw.reset();
>> 	sw.start();
>> 	for(double i = 0; i < 10000000; i++) xxx += sin(PI*i);
>> 	t = sw.peek().msecs;
>> 	writeln(t);
>> 	sw.stop();
>> }
>
> What you are doing wrong is that xxx is never used... So DMD 
> removes it altogether?
>
> If you add writeln(xxx) after the second loop as well, then 
> maybe the measurements make more sense?
>
> Ola.

maybe, I thought I put in the writeln(xxx) in there to do that... 
yeah, looking at my code it's there, so that is not the problem.

This is the code I had,

	import std.datetime;
	double xxx = 0;
	StopWatch sw;
     sw.start();
	for(double i = 0; i < 10000000; i++) xxx += sinTab(i);
	auto t = sw.peek().msecs;
	writeln(t);
	sw.stop();
	writeln(xxx);
	
	xxx = 0;
	sw.reset();
	sw.start();
	for(double i = 0; i < 10000000; i++) xxx += sin(PI*i);
	t = sw.peek().msecs;
	writeln(t);
	sw.stop();
	writeln(xxx);

Either I modified it later or deleted one of those lines or the 
editor screwed up when I pasted it... I don't recall touching the 
code afterwards.

Either way, sin it's still twice as fast. Also, in the code the 
sinTab version is missing the writeln so it would have been 
faster.. so it is not being optimized out.