jprobe of jcs

[Aaron Smuts <aa...@verizon.net>]

Some profiling of JCS.  The relative timing is ok, but it cannot be
taken as absolute time since it is running inside JProbe.

I'll see if I can get it some better data.

The LinkedQueue is significant.

All the bottlenecks are in disk io.

Aaron

> -----Original Message-----
> From: dlr@despot.finemaltcoding.com
[mailto:dlr@despot.finemaltcoding.com]
> On Behalf Of Daniel Rall
> Sent: Thursday, February 21, 2002 11:20 PM
> To: kaz@kazmier.com
> Cc: Turbine Developers List
> Subject: Re: [DOCUMENTATION] Stratum / JCS Documentation
> 
> Cool, thanks Pete.
> 
> "Pete Kazmier" <pe...@kazmier.com> writes:
> 
> > Dan,
> >
> > There are quite a few xdocs up on the Stratum site about JCS.  I
believe
> > Aaron's most recent addition, "Basic JCS Configuration", was
prompted by
> > this very thread.
> >
> > Thanks,
> > Pete
> >
> > On Thu, Feb 21, 2002 at 09:17:38AM -0800, Daniel Rall wrote:
> > > Anyone get this good stuff into an xdoc yet?
> 
> --
> To unsubscribe, e-mail:   <mailto:turbine-dev-
> unsubscribe@jakarta.apache.org>
> For additional commands, e-mail: <mailto:turbine-dev-
> help@jakarta.apache.org>

Re: jprobe of jcs

[James Taylor <jt...@4lane.com>]

On Fri, 2002-02-22 at 00:28, Aaron Smuts wrote:
> Some profiling of JCS.  The relative timing is ok, but it cannot be
> taken as absolute time since it is running inside JProbe.
> 
> I'll see if I can get it some better data.
> 
> The LinkedQueue is significant.

Is it? Can you elaborate on your interpretation of the data. It'd be
great to see similar profiling of the version with the JGLQueue.
LinkedQueue looks pretty quick to me -- most of the time in the take
operation is from waiting for objects to become available. I'm not
entirely sure how to read the profiling output though.

> All the bottlenecks are in disk io.

NIO!

-- jt



--
To unsubscribe, e-mail:   <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>