Numerical Computing with Modern FortranRichard J. Hanson and Tim Hopkins |
Chapter 11: Defined Operations for Sparse Matrix Solutions
Source Code:
The code associated with this chapter requires version 4.3 of the SuperLU package.
There is a single executable that may be built using the makefile.
- testSLUInterop uses set_precision.f90, sluInterop.f90, sortElements.f90, sparseAssign.f90, sparseOps.f90, sparseSort.f90, sparseTypes.f90, fortran_double_call_SuperLU.c, and SuperLU v4.3 as a library.
Sample output from testSLUInterop
This used the Intel compiler, XE 14.0.1.139, and Intel MKL for routine DNRM2.
The SuperLU factorization completed with N = 5 The nonzero values of L, U and their sum = 11 11 22 Memory Usage (KBytes) for LU step = 2.720000E-01 Total Usage (KBytes) for all terms = 1.412000E+00 Test Status Description 1 Passed H-B Matrix (.ip.) b(:) Agrees with known solution. 2 Passed b(:) (.pi.) H-B Matrix Agrees with known solution.