User:Edu159: Difference between revisions

Jump to navigation Jump to search
5,357 bytes added ,  28 March 2014
 
(21 intermediate revisions by 2 users not shown)
Line 17: Line 17:


== C: Contact ==
== C: Contact ==
'''* Please state the (unique and identical where possible) nick you use on IRC and any other communication channel related to Octave. ''We really want unique nicks. You might want to bold it.'''''
*''' Please state the (unique and identical where possible) nick you use on IRC and any other communication channel related to Octave. ''We really want unique nicks. You might want to bold it.'''''
nick: edu159  
nick: edu159  


'''* Which time zone do you live in? Will that change over GSoC duration? ''Perhaps DST adjustment or a relocation. Note that both UTC and GMT are not aware of daylight saving time! Please state UTC+x or -x.'''''
*''' Which time zone do you live in? Will that change over GSoC duration? ''Perhaps DST adjustment or a relocation. Note that both UTC and GMT are not aware of daylight saving time! Please state UTC+x or -x.'''''


Time zone: UTC+0. It won't change during GSoC.
Time zone: UTC+0. It won't change during GSoC.


'''* Please state the timeframe (in UTC+0) when you feel most comfortable working during GSoC. Where are your time buffers? ''Example: I usually code around 9.00 to 18.00 and could also try to start earlier (~7.00) for few days ;-)'''''
*''' Please state the timeframe (in UTC+0) when you feel most comfortable working during GSoC. Where are your time buffers? ''Example: I usually code around 9.00 to 18.00 and could also try to start earlier (~7.00) for few days ;-)'''''


I can code from 7:00-13:00 in May, June and first half of July from Monday to Thursday. On Fridays, Saturdays and Sundays I have almost the whole day free for coding. In the second half of July and August I have no timeframe restrictions.
I can code from 7:00-13:00 in May, June and first half of July from Monday to Thursday. On Fridays, Saturdays and Sundays I have almost the whole day free for coding. In the second half of July and August I have no timeframe restrictions.
Line 37: Line 37:
I have only that kind of experience while doing some personal projects with a friend of mine, so I have not experience with such a large project and free software community.
I have only that kind of experience while doing some personal projects with a friend of mine, so I have not experience with such a large project and free software community.
* '''Please describe the biggest project you have written code for and what you learned by doing so. Also describe your role in that project over time.'''
* '''Please describe the biggest project you have written code for and what you learned by doing so. Also describe your role in that project over time.'''
I have dealt with projects with around 2.5-3k lines of code, not more. I don't have a extraordinary successful story about a project I have been in although I have been in lots of small ones. I worked the last year while in college with a friend on a website for informal job seeking done in PHP+MySQL (around 3K lines of PHP code). We even sent the idea to Ycombinator to get funding. Unfortunately we were not selected and the idea was not at good as we thought. I learned a lot with that project, particularly combining both college and a project at the same time, working regularly and establishing deadlines.
I have dealt with projects with around 2.5-3k lines of code, not more. I don't have a extraordinary successful story about a project I have been in although I have been in lots of small ones. I worked the last year while in college with a friend on a website for informal job seeking done in PHP+MySQL (around 3K lines of PHP code). We even sent the idea to Ycombinator to get funding. Unfortunately we were not selected and the idea was not as good as we thought. I learned a lot with that project, particularly combining both college and a project at the same time, working regularly and establishing deadlines.
* '''Please state the commits and patches you already contributed to Octave. ''This question (one of the most important parts by the way) is the only part of your application our wiki admins will edit for you even after the application deadline. Code sometimes speaks louder than many words do.'''''
* '''Please state the commits and patches you already contributed to Octave. ''This question (one of the most important parts by the way) is the only part of your application our wiki admins will edit for you even after the application deadline. Code sometimes speaks louder than many words do.'''''


*Contributions to Octave:
*Contributions to Octave:
** Patch submitted for bug [http://savannah.gnu.org/bugs/?41839 #41839] (m-file)
** Patch submitted for bug [http://savannah.gnu.org/bugs/?41839 #41839].(m-file) '''UPDATE:''' The patch has been accepted.
** Patch submitted for bug [http://savannah.gnu.org/bugs/?41796 #41796] (c++)
** Patch submitted for bug [http://savannah.gnu.org/bugs/?41796 #41796].(c++)
*Others:
** Code that implements partial functionality of '''ilu''' function("no-fill option") : [https://db.tt/hzpWkTkq file]


== F: Feeling fine ==
== F: Feeling fine ==
Line 75: Line 77:
I like programming on Linux.
I like programming on Linux.
* '''If you have access to more than one, please state them and the conditions under which you are granted this access.'''
* '''If you have access to more than one, please state them and the conditions under which you are granted this access.'''
I have windows (native) and mac virtualized.
I own also a Windows native installation.
* '''Please estimate an average time per day you will be able to ''(if separated)'' access'''
* '''Please estimate an average time per day you will be able to ''(if separated)'' access'''
** an internet connection --> ~18 hours a day
** an internet connection --> ~18 hours a day
Line 95: Line 97:


== Y: Your task ==
== Y: Your task ==
My intention is upgrading  some functions related with sparse matrices  so they become compliant with Matlab  and implement others that are not present in Octave right now. Next
The functions are '''ichol''', '''ilu''', '''sprand''', '''sprandn''', '''sprandsym''', '''minres''' and '''lsqr'''.


* Approach:
'''Introduction'''
**'''ilu:''' That function has a big chunk of options and the last year was almost implemented by Kai Torben as his GSOC project. He interfaced Octave with ITSOL/ZITSOL libraries but in the end there were some issues with that approach:
------------
My intention is to implement the missing '''ilu''' and '''ichol''' functions and to continue last year's GSoC project developed by Kai Torben. The goal is the complete integration in the core of Octave. Those functions will have a direct impact on quite a few functions that solve sparse linear systems as ilu and ichol are known to be good algorithms for generating preconditioners. Some of the functions that will take advantage of them are pcg, bicg or gmres among others. Bottom line, performance of sparse linear system solvers in Octave will be raised.


    1) ILUTP algorithm did not work for him
    2) He had to patch the library to get things work!
    3) modified versions of algorithms ("milu" option) were not implemented in the libraries
    4) That "ugly" scenario lead to finally not being able to include ITSOL as a dependency with Octave. Bottom line, the integration of the function with the development repository could not be achieved.


'''Approach'''
----------------


*'''ilu:''' This function has a big chunk of options and the last year was almost implemented by Kai despite of the fact that he could not finally integrate it into Octave . His approach was interfacing Octave with ITSOL/ZITSOL libraries but in the end there were some issues with that direction:
:#ILUTP algorithm did not work for him
:#He had to patch the library to get things work. Not nice for achieve integration.
:#modified versions of algorithms ("milu" option) were not implemented in the libraries
:#That "ugly" scenario lead to finally not being able to include ITSOL as a dependency with Octave. Bottom line, the integration of the function with the development\n repository could not be achieved.


::I have been in contact with Kai by mail and agrees that writing from scratch all the functions needed as oct-files (ILUTP, ILU0, ILUC and ILUT) would be a harder but safer way to go if integration want to be achieved. This way no dependencies are needed to be added and the overhead of translating the data from Octave to ITSOL and vice verse is eliminated. Algorithms will be taken from Yousef Saad's  book "Iterative methods for sparse linear systems Ed. 2". Moreover, I can use some of the code that Kai wrote, mostly the tests, documentation and the m-file "ilu.m" that glue together all the functions. ITSOL source code is also a good place to look for some help.


* Please provide a rough estimated timeline for your work on the task. ''This should include the GSoC midterms and personal commitments like exams or vacation ("non-coding time"). Optionally include two or three milestones you expect.''
::I have implemented the ILU0 algorithm so far and benchmarked it against Matlab and Kai's last GsOC version(using ITSOL). The performance is great. You can check the code and see a table with the execution times in a [http://edu159-gsoc2014.blogspot.com.es/2014/03/ilu0-implementation.html blog] I have created to inform about my achievements in the project. [https://db.tt/hzpWkTkq Here] is the link to the source code I have written so far. I has not got any doc-strings nor tests, but is functional.
[[Category: Summer of Code]]




Under construction....
*'''ichol:''' In this case things should be easier at first glance. Kai  implemented the functions related with ichol from Fortran prototypes and they work as expected. In theory would be only necessary to code the complex versions and the modified versions of the algorithms. On the other hand, there is an issue here with licenses I did not know at first and Kai pointed me out (see [http://edu159-gsoc2014.blogspot.com.es/2014/03/introducing-myself.html#bc_0_1B here]). Authors should be contact in order to ask for a permission, but if they are not in favor of giving to us then functions will be implemented from scracth.
 
 
:The road map of what and how I want to do the project is clear to me, but I don't know for sure if it will be enough for the GSoC period. Because I am concerned about that there are other functions that I would like to implement if it would be necessary like '''lsqr''' and '''minres''', both highly related with ichol and ilu. I have already done some search and found that this website [[http://www.stanford.edu/group/SOL/software/lsqr/ lsqr]] [[http://www.stanford.edu/group/SOL/software/minres/ minres]]. The website is from the people that wrote the papers given as references in Matlab documentation. In the website there are several codes that can be used. I have mailed professor Michael Saunders about adapting them into Octave versions and he answered me that I am welcome to do while I respect the license (CPL or BSD licenses). He claimed that they are very unrestrictive but I've been told that they are not compatible with GPL3. I will need some insights about that if I happen to have time for implementing them.
 
 
'''Project structure'''
-------------------------
 
These are the main files that will be needed for the project:
 
* ilu.m -- Act as a wrapper for the rest of ilu functions
** ilu0.cc -- Implements ilu decomposition zero level of fill
** ilut.cc -- Implements ilu decomposition with threshold
** iluc.cc -- Implements Crout version of ilu decomposition
** ilutp.cc-- Implements the ilu decomposition with threshold and pivoting
 
*ichol.m -- Act as a wraper for the rest of functions
** ichol0.cc -- Implements ichol decomposition with zero level of fill
** icholt.cc -- Implements ichol decomposition with threshold
** icholt0jp.cc -- Implements ichol decomposition following Jones and Plassmann strategy
 
NOTE: All the algorithms could be used with real and complex numbers and "modifified versions" will be implemented as well.
 
 
'''Estimated timeline:'''
-------------------------
 
 
*'''FIRST PERIOD:'''
::'''19 May-20 June:'''  Implement ilu related functions. In that order ilu0, ilut, iluc , ilutp.
::'''20-25  June:'''  Write ilut, iluc, ilutp automated tests, documentation and benchmarking.
::'''27    June:'''  '''(Millstone 1)''' ilu.m is fully functional and ready to be integrated with Octave core.
::'''27    June:'''  Mid-term evaluation
 
*'''SECOND PERIOD:'''
::'''28  June-2 July:''' Strategy for ichol will be fully established in those days. Better to plan wisely before code.
::'''2 July-5 August:''' Implement ichol related functions. In that order ichol0, icholt. ichol0jp if there is time left.
::'''5-10    August:''' Write automated tests, documentation and benchmarking for those functions.
::'''10    August:''' '''(Millstone 2)''' ichol.m is fully functional and ready for integration.
::'''10-18 August:'''  Buffer week for any unexpected situation or minor changes that should be done. If there is time left, then I will work on minres and lsqr functions.
 
 
'''Note:''' I set schedule starting on May 19th but I would like to start coding since I know I am selected (22 April), so maybe goals would be reached before stated above. So that dates should be taken as top limits. In June 3 college exams are scheduled that still not have fixed dates. I am not worried about them since I have a lot of time for studying.
77

edits

Navigation menu