Quick latency testing question..

[Samgetsmoney]

ok at the beginning of the boolean i have
long start = System.currentTimeMillis();
and at the end i have
long end = System.currentTimeMillis();
long latency = end - start;



How do i make it more precise?
Like System.currentTimeNanos()?
lol


this is what i mean
pastebin*/m21d2c9af

where the * is put .com

Thanks

[Fubarable]

why do you need timing more precise than milliseconds? Perhaps instead you need to correct your design.

[Samgetsmoney]

Well then help with my design..
Because latency just equals 0 all the time..

[Fubarable]

Well then help with my design

I'd love to do just that if I had any inkling what the heck you're trying to do. I must have left my mind-reading hat back at work.

You may wish to provide more detail. A great site to help you formulate your questions so that they are more easily answered can be found here: How To Ask Questions The Smart Way

[Samgetsmoney]

And a quick question why when u do
Thread.sleep(NUMBER);
it never actually sleeps for that long?

[Samgetsmoney]

look at the part where it says pastebin thats what i mean

[MK12]

The pastebin thing just has the start then end then calculates latency. obviously it will be zero because you're not doing something in between.. Do something that takes a long time in the middle, like printing out 500 pascal's triangles that each have 100 rows:D.
-MK12

[Samgetsmoney]

yea but when the method is running 100's of times every half second the little bit of latency counts..

And there is code in the middle the commented out stuff was just an example...

So i read up on benchmarks but dont fully understand them..

[CJSLMAN]

A couple of articles that discuss Java benchmarking that are referenced a lot:

Robust Java benchmarking, Part 1: Issues
Robust Java benchmarking, Part 2: Statistics and solutions

Luck,
CJSL