[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ale] Processor usage

Unless the java code was compiled to be multi-threaded, thread safe and teh
entire applicaiton designed to support parallel processing, nope. (welcome
to my world :-)

That said, you could split the video job into 24 similar sized chunks and
start a process for each one.

When you see the utilization spike, run top and then press '1' to see the
per processor breakdown. If you see cpu0 running hard then falling off
while cpu1 ramps up (repeat for other cpus), you're also getting clobbered
by thread-migration overhead stuff. It's possible to do some cpu-affinity
work and pin a process to a specific cpu so it will no longer move around
and waste cycles just refilling the cache. Look up the tool "taskset" for
processor affinity abilities. On CentOS 6.5 it's in the util-linux-ng

On Thu, Feb 27, 2014 at 4:49 PM, Adrya Stembridge <
adrya.stembridge at gmail.com> wrote:

> A colleague and I were looking at some ontology data yesterday using java
> via command line.  I noticed that CPU utilization averaged 99% on a system
> with 24 cores.  The data processing is intense and time consuming.  We're
> wondering if this could be sped up by forcing the machine to use more CPU
> resources.    In the past, I've seen utilization reach around 1500% while
> encoding video.
> Is it possible (or safe) to instruct or force java to use more cores in
> processing our data, or am I looking at this the wrong way?
> as
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo

James P. Kinney III

Every time you stop a school, you will have to build a jail. What you gain
at one end you lose at the other. It's like feeding a dog on his own tail.
It won't fatten the dog.
- Speech 11/23/1900 Mark Twain

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20140227/ac30257d/attachment.html>