Are there any software Packages/Libraries which help in implementation of Machine Learning Algorithms to make use of multiple core or multi processors machine.

asked Jun 30 '10 at 23:57

anon's gravatar image

anon
1473610

reverted Jul 01 '10 at 18:31

Joseph%20Turian's gravatar image

Joseph Turian ♦♦
579051125146

Has anyone had success parallelizing an online learning algorithm. For instance, MIRA or CW learning?

(Jul 01 '10 at 00:06) aria42

reverted, but leaving the question closed.

(Jul 01 '10 at 18:31) Joseph Turian ♦♦

The question has been closed for the following reason "This question is too general, and seems like a solution in search of a particular problem. Why this family of ML algorithms? And why a multi core solution instead of a parallel one?" by Joseph Turian Jul 01 '10 at 01:03


One Answer:

Most interesting BLAS implementations can take advantage of multiple cores and can achieve close to linear speed-up for operations such as matrix multiplications. If your Machine Learning algorithms/library does matrix multiplications and can call BLAS api functions, then you're good to go.

answered Jul 01 '10 at 00:06

Dumitru%20Erhan's gravatar image

Dumitru Erhan
12169

powered by OSQA

User submitted content is under Creative Commons: Attribution - Share Alike; Other things copyright (C) 2010, MetaOptimize LLC.