From benson@mcs.anl.gov Thu Mar 27 14:49:20 2003
Date: Thu, 27 Mar 2003 14:20:50 -0600 (CST)
From: Steve Benson <benson@mcs.anl.gov>
To: Hong Zhang <hzhang@mcs.anl.gov>
Subject: mats and vec.


Hi Hong,

   These matrices are in PETSc binary format.  I have also enclosed a 
the driver that I used to test them.   The diagonal jacobi preconditioner
seems to work, although slowly.

   The came from the TAO example called Ginzburgh-Landau, from the 
MINPACK-2 test suite.

   The same application can create more matrices.  If you would like
to have more, let me know.

Steve

To solve system using the driver, compile it, then:
sles_driver -vecfile vec24.b -matfile mat24.b -ksp_max_it 1000 
-ksp_monitor 




    [ Part 2, "Big matrix"  Application/OCTET-STREAM (Name: "mat24.b")  ]
    [ 1.3MB. ]
    [ Unable to print this part. ]


    [ Part 3, "Big RHS"  Application/OCTET-STREAM (Name: "vec24.b")  ]
    [ 25KB. ]
    [ Unable to print this part. ]


    [ Part 4, "Smaller matrix"  Application/OCTET-STREAM (Name: "mat6.b")  ]
    [ 86KB. ]
    [ Unable to print this part. ]


    [ Part 5, "smaller RHS"  Application/OCTET-STREAM (Name: "vec6.b")  ]
    [ 1.5KB. ]
    [ Unable to print this part. ]


    [ Part 6, "My C driver"  Text/PLAIN (Name: "sles_driver.c")  44 ]
    [ lines. ]
    [ Unable to print this part. ]


    [ Part 7, "My makefile"  Text/PLAIN (Name: "makefile")  13 lines. ]
    [ Unable to print this part. ]