Some of the sequential classifiers use multiple cores for training even without Hadoop.
On Sun, Oct 16, 2011 at 3:20 AM, Sean Owen <[email protected]> wrote: > I think what you're really asking is, can I run a Mahout job on Hadoop > using 1 worker, 2 workers, 4 workers? Yes. But this is a Hadoop > question, not Mahout. > > I'd also say that you are probably going to find these things are > bottlenecked on I/O, not CPU. With 8 Hadoop workers on one machine, I > doubt it will be using 100% CPU. > > > On Sun, Oct 16, 2011 at 9:36 AM, d_col <[email protected]> wrote: > > Hi, I am a beginner in Mahout and I would like to know if it possible to > run > > an algorithm in solely one of the cores of my (8-core) cpu then 2 cores, > > then 4 and so on... and then see the improvement in execution time. > > > > My real question is: is there a formal way of taking advantage of > multiple > > cores the same way you take advantage of multiple cpus? > > > > Thanks a lot! > > > > -- > > View this message in context: > http://lucene.472066.n3.nabble.com/Can-Mahout-work-with-cores-instead-of-cpus-tp3425604p3425604.html > > Sent from the Mahout User List mailing list archive at Nabble.com. > > >
