[Mono-list] tiny benchmark of current mono clr, java clr, ms clr

[Jonathan Pryor]

Much as I'd hate to ding your benchmark, but there's one minor problem
with it: you're not really testing the performance of the different
runtime environments, as much as you're testing each platforms console
support.  I've found Windows console output to be SLOW.  Phenomenally
slow.  Monumentally slow.  Even with optimized, natively-compiled C
apps.  Under Windows, I feel like I can see each line being drawn (at
least on my machines); granted, it's a very fast line draw, but I can
still make out it out.  On Linux, the console is so fast I typically
can't see the console redraw.

This is why I'm not suprised that the fastest Windows time is slower
than the worst Linux time.

Actually, I'm surprised that Mono on Windows ran as fast as it did,
given my history with the Windows console.

A better comparison would avoid the console, and just stick with
function invocation (for example).  Unfortunately, I can't think of any
way to keep functions from being inlined by .NET, leading to
ridiculously fast execution times due to inlining.

Such is the difficulty with benchmarks.

 - Jon

On Sun, 2002-11-10 at 14:42, Daniel Mettler wrote:
> hia,
> 
> i just made a *tiny* (very limited significance) benchmark test
> of the current versions of some runtime environments on linux and
> windows.
> 
> http://www.icu.unizh.ch/~mettlerd/variouspapers/mono_bench/
> 
> actually it's not noteworthy. on the other hand posting it here
> does not hurt either ;)
> 
> regards
> 
> dan
> 
> ps. miguel, i'm looking forward to attending your talk at the
> university of zurich (http://www.icu.unizh.ch/monotalk/, today:
> server downtime)
> 
> -- 
>       ...::: Daniel Mettler | http://www.numlock.ch :::....
> 
> 
> 
> _______________________________________________
> Mono-list maillist  -  Mono-list@ximian.com
> http://lists.ximian.com/mailman/listinfo/mono-list