[Discuss] Comparison of Java performance on Ubuntu Linux versus Windows Vista

[Cy Schubert]

In message <494D45EB.8030306 at gmail.com>, Adam Parkin writes:
> Alan W. Irwin wrote:
> > Well, Vista versus XP is not too interesting a topic to me because I don't
> > care very much about windows details.  
> 
> Well it's not so much XP vs Vista I was getting at (that doesn't 
> particularly interest me either), but rather Sun JVM vs Windows OS that 
> was the question for me.  Like I said, if the slowdown occurs on one 
> version of the OS and not another when the JVM is the same, then it's 
> likely the OS, if it doesn't, then it could very well be the JVM that's 
> the issue.  And if that's the case, that would beg the question why is 
> the Windows Sun JVM slower than the Linux version?  Is it because of 
> architectural choices in the design of the OS that allow the Linux 
> version to be faster, or is it because Sun did a poor job implementing 
> the Windows version of their JVM (or is it something else entirely?)

Not necessarily. The compiler and optimisation would have a greater impact 
on performance than the OS. How much kernel time does the JVM use v.s. user 
time. The other factor is shared libraries/DLLs. This would be difficult 
without the use of profiling libraries or a tool such as DTrace.

> 
> I guess I just don't like benchmark results without explanations. :p


-- 
Cheers,
Cy Schubert <Cy.Schubert at komquats.com>
FreeBSD UNIX:  <cy at FreeBSD.org>   Web:  http://www.FreeBSD.org

			e**(i*pi)+1=0
>