Java multi CPU capabilities
William Muriithi
william.muriithi-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org
Thu Aug 29 12:25:09 UTC 2013
On Aug 29, 2013 12:08 AM, "D. Hugh Redelmeier" <hugh-pmF8o41NoarQT0dZR+AlfA at public.gmane.org> wrote:
>
> I don't use Java. Ted doesn't use Java. I think Lennart must have
> used Java but has been trying to expunge the memory ever since.
>
> So what we say is unreliable in that sense. But we each know a bunch
> of things about coumputer systems in general.
>
Okay, I understand.
> | > What do you mean by "natively"?
> | >
> | Apparently Java has two types of switches, the green and native switch.
The
> | native can only use one kernel thread which essentially mean one CPU.
>
> I don't know why switches would be green or native. I hope that this
> is an irrelevant detail.
>
Hope so too. I may read more about it when I get a chance purely for
curiosity.
>
> | Totally see why I sound lost. I possibly don't what I am taking about
but I
> | have checked cacti graphs at work and all seem to max up on one CPU.
>
> What do you mean by a cacti graph? A pretty presentation of
> performance monitoring? <http://www.cacti.net/>
>
Petty much. Have set it up to monitor all kind of statistics. I usually
look at it when a system get a problem in an attempt to find the root
cause. It current monitor all CPU in a system
> | Google a bit and you will see a lot of people having that problem. Java
> | can use multiple CPU together for garbage collection but the application
> | side, its only one CPU as far as I can tell
>
> What's "that problem"? Wack of exploitation of more than one core?
>
That was what initially lead me to investigate a bit. Not really a problem
in that nothing is broken but a curious part of me wanted to make sense of
the observations
> As I google, I find more evidence of Java concurrency exploiting
> "multicore". See
> <
http://docs.oracle.com/javase/tutorial/essential/concurrency/highlevel.html>
>
Interesting, I admit I am missing something here. Its really possible we
have some setting wrong. If I ever find a solution, will update the list.
>
> | From: William Muriithi <william.muriithi-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org>
>
> | But how do you achieve that if your application is restricted without a
> | container that is only single kernel threaded?
>
> What do you mean by "container"? That has several quite different
> meanings in the Java world.
>
Hmm, I didn't know that. Will read on all the meaning and clarify.
> As it is, I cannot understand your question.
> --
Okay my bad. Ya, with the links you shared, its more likely something
inherent to our setup than the Java. Or even a misunderstanding of the
statistics I correct from the systems.
William
> The Toronto Linux Users Group. Meetings: http://gtalug.org/
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gtalug.org/pipermail/legacy/attachments/20130829/3eb9e65c/attachment.html>
More information about the Legacy
mailing list