Browse by Tags

Matrix data structure optimization: Nearly 50% perf gain

Some time ago I did some very basic performance analysis of solving linear equation systems with Iridium. Since then we decided to rewrite the Matrix class to use jagged arrays instead of rectangular ones, see this forum discussion . There already was a discussion about that issue some long time ago...

Posted to Math.NET Project News (Weblog) by Christoph Rüegg on 10-14-2007
Re: Matrix class performance

I've submitted some changes on the matrix class (and dependencies) today, so it now uses jagged arrays, as you suggested. Indeed, I observed nearly 50% perf improvement on the included small Solve perf analysis project. You may want to have a look at the changes first before updating your local SVN...

Posted to Math.NET Mathematics Library (Forum) by Christoph Rüegg on 10-14-2007
Re: Matrix class performance

Hi, Thanks very much for your analysis. Your numbers speak a clear language. Interestingly I had just talked with another user about using jagged arrays instead of the rectangular ones some days ago... Looks like we'll have a new matrix base in the next iridium release :) Thanks, Chris PS: Are you...

Posted to Math.NET Mathematics Library (Forum) by Christoph Rüegg on 10-08-2007
Matrix class performance

I have done some qualitative tests comparing different array types as storage for a Matrix class and have found that the rectangular array is the slowest possible way to implement the Matrix class. The other two options that I considered where jagged arrays (double[][]) and linear arrays(double[]) where...

Posted to Math.NET Mathematics Library (Forum) by gengelb on 10-04-2007
very large matrix to be inverted

dear all, I came across math.net during a struggle to create and invert a very large matrix ( minimum= 5000x5000 and growing to 100000 x 100000) although some figures shown on the website look promissing, i am unable tocreate a matrix larger then 4000x4000 ad even that is with trouble. I keep gtting...

Posted to Math.NET Mathematics Library (Forum) by leon on 09-17-2007
Thread Scheduler

I think that it would be a good Idea to realize a kind of thread scheduler that provides one (or two if hyperthreading is enabled) thread(s) per core. If a calculation is able to get parallelized, like the calculation of matrix determinants or the calculus of integrals, the thread scheduler should get...

Posted to Math.NET Mathematics Library (Forum) by MovGP0 on 05-27-2007
Iridium Performance Analysis

I started performing some very basic performance analysis of interesting parts of the Math.NET Iridium (Numerics) Library, motivated by a forum post . Machine Configuration I tested it on the following computers: A : Desktop, Intel P4, 3.4 GHz , XP SP2 B : Notebook, Intel Core 2 Duo, 2 GHz , XP MCE SP2...

Posted to Math.NET Project News (Weblog) by Christoph Rüegg on 04-15-2007
Re: Limits of Matrix class

Yes, it's a 3.4 GHz P4 desktop I've just tested it also on my notebook (2GHz): Debug Code: 100x100: 5 ms 200x200: 40 ms 330x330: 160 ms 1000x1000: 4000 ms Release Code: 100x100: 1 ms 200x200: 20 ms 330x330: 70 ms 1000x1000: 1900-1950 ms Apparently the code built in release mode is about twice...

Posted to Math.NET Mathematics Library (Forum) by Christoph Rüegg on 04-01-2007

Page 1 of 1 (8 items)