Running Sparse Solvers!

32 views
Skip to first unread message

Esmail Abdul Fattah

unread,
Jul 12, 2023, 3:36:31 PM7/12/23
to up...@googlegroups.com
Dear Team!!

I am very interested in testing the sparse solver (from Sympack) with some sparse matrices generated from Statistical applications. I have succeeded in installing Sympack and UPCXX, but not running a simple example that solves a system of equations based on sparse matrices. Could you please provide a running .cpp example for a sparse solver?

The online example from https://github.com/symPACK/symPACK :

#first download the input matrix (we are using nasa2146 here)
wget https://sparse.tamu.edu/RB/Nasa/nasa2146.tar.gz
#extract the matrix
tar xzf nasa2146.tar.gz

#run sympack
upcxx-run -n 4 -- ./run_sympack -in nasa2146/nasa2146.rb -ordering MMD -nrhs 1

is not very helpful.

Best Regards,
Esmail 


This message and its contents, including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.

Julian Bellavita

unread,
Jul 12, 2023, 5:40:45 PM7/12/23
to UPC++
Hi there.

Are you trying to set up symPACK in a program separate from the provided driver in driver/run_sympack2D.cpp, or are you trying to run this driver and running into an error?

If it is the former, and you are looking for an example of how to use symPACK in a separate application, I would suggest looking at driver/run_sympack2D.cpp. That file demonstrates the complete process of initializing symPACK and performing all the phases of sparse cholesky factorization. It should hopefully serve as a decent example of how to incorporate symPACK into your application. 

If it is the latter, and you are attempting to run the example provided in the README but are running into an error, could you please send the error message? 

Thank you, and apologies if I am misunderstanding something about your question!

-- Julian

Reply all
Reply to author
Forward
0 new messages