RE: about performance statistics of PCA.dml

2017-07-21 Thread arijit chakraborty
Thanks a lot Niketan! This was a great help! Sent from Mail for Windows 10 From: Niketan Pansare Sent: Saturday, July 22, 2017 1:04 AM To: dev@systemml.apache.org Subject: RE: about performa

RE: about performance statistics of PCA.dml

2017-07-21 Thread Niketan Pansare
Yes, please remove the ".template" suffix, place SystemML-config.xml in the current directory and set the property systemml.stats.finegrained to true: https://github.com/apache/systemml/blob/master/conf/SystemML-config.xml.template#L73 Thanks, Niketan Pansare IBM Almaden Research Center E-mail: n

RE: about performance statistics of PCA.dml

2017-07-21 Thread arijit chakraborty
Hi Niketan, Sorry to get in between you and Janardhan’s chat. Actually I’m having issue in improving the performance of my system. You suggested me how to get some stats and I’ve incorporated it. But things after “Heavy hitter instructions:”, seems very useful for me to debug where the system i

Re: about performance statistics of PCA.dml

2017-07-21 Thread Janardhan Pulivarthi
Thanks a lot niketan. On Fri, Jul 21, 2017 at 11:28 PM, Niketan Pansare wrote: > Hi Janardhan, > > You can get instruction-level statistics with the commit > https://github.com/apache/systemml/commit/648eb21d66f9cd8727090cdf950986 > 765a7e6ee8: > SystemML Statistics: > Total elapsed time: 18.956

Re: about performance statistics of PCA.dml

2017-07-21 Thread Niketan Pansare
Hi Janardhan, You can get instruction-level statistics with the commit https://github.com/apache/systemml/commit/648eb21d66f9cd8727090cdf950986765a7e6ee8 : SystemML Statistics: Total elapsed time: 18.956 sec. Total compilation time: 1.924 sec. Total execution time: 1

about performance statistics of PCA.dml

2017-07-21 Thread Janardhan Pulivarthi
Hi Mike, I'd like to know how much expensive this critical code is C = (t(A) %*% A)/(N-1) - (N/(N-1))*t(mu) %*% mu; (at https://github.com/apache/systemml/blob/master/scripts/algorithms/PCA.dml#L81) in the SPARK setting given 1. 60Kx700 input for A 2. For a datasize of 28 MB with 100 con