[Shootout-list] fannkuch (timer resolution; HZ=1000?)

[Isaac Gouy]

--- Brent Fulgham <bfulg@pacbell.net> wrote: 
> --- Jon Harrop <jon@ffconsultancy.com> wrote:
> > Two bigger problems are that most modern language
> > will add much bigger random noise and a bias. OCaml 
> > running times often vary by 0.1s and I've heard
> > (from Microsoft) that C# does poorly in the shootout
> > because it has a big startup time.
> 
> C# doesn't do so poorly.  It beats the pants off of
> Java (meaning Sun's HotSpot version) in almost all
> tests.

Of course, we're talking about Mono not Microsoft C#.


And anybody making the comparison will see that Sun Server Java takes
2.4x longer than Mono C# to startup.

http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=java&lang2=csharp&sort=fullcpu#ratio


And this is such old news that the FAQ links to the crude measure of
startup from Doug Bagley's original shootout

http://shootout.alioth.debian.org/great/benchmark.php?test=hello&lang=all&sort=fullcpu

C# takes at least 30x longer than gcc to startup, Java Server takes at
least 70x longer than gcc to startup.


> > I thought it would be nice to do those Skaller plots
> > using only tests which ran for >0.1s or even >1s but
> > there is virtually no such data on the shootout 
> > (i.e. the plots are all blank).
> 
> So, we should probably up the 'N' on most tests until
> we can get about 0.5 seconds or so on the fastest
> performer.
> 
> Obviously, this causes problems for the slow
> languages (since they will most likely time out now).

We actually reduced the max time for nbody within the last week,
because  I'd made it too large.

As there are Shootout contributors who work in Scheme and Tcl and
Python, we've been balancing the data range on each test - so that
those languages never timeout at min N (but occasionally timeout at max
N), and so that optimized compiled languages get enough run-time at max
N.


Amazingly this allows Python to be compared to Ruby

http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=python&lang2=ruby&sort=fullcpu

and Intel C to be compared to GCC

http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=icc&lang2=gcc&sort=fullcpu


Bagley called it the Great Computer Programming Language Shootout
because it compared a great many language implementations - not just
the 5 most optimised compilers. 



		
__________________________________ 
Do you Yahoo!? 
Read only the mail you want - Yahoo! Mail SpamGuard. 
http://promotions.yahoo.com/new_mail