next up previous contents
Next: Notes on Use Up: NSPCG User's Guide Previous: Using Multicolor Orderings

   
Error Conditions

The NSPCG package assigns to an integer variable IER certain values to indicate error conditions (or lack of them) in the program's execution. Positive values are assigned to warnings, and negative values are assigned to fatal errors. The values IER can be assigned and the corresponding meanings are indicated in the following table.


 
Table 8: Meanings of Assigned IER Values
IER Meaning
0 No error detected
-1 Nonpositive matrix size N
-2 Insufficient real workspace
-3 Insufficient integer workspace
-4 Nonpositive diagonal element
-5 Nonexistent diagonal element
-6 A is not positive definite
-7 Q is not positive definite
-8 Cannot permute matrix as requested
-9 MDIM is not large enough to allow expansion of matrix
-10 Inadmissible parameter encountered
-11 Incorrect storage mode for block method
-12 Zero pivot encountered in factorization
-13 Breakdown in direction vector calculation
-14 Breakdown in attempt to perform rotation
-15 Breakdown in iterate calculation
-16 Unimplemented combination of parameters
-18 Unable to perform eigenvalue estimation
1 Failure to converge in ITMAX iterations
2 ZETA too small - reset to 500*SRELPR
3 ZBRENT failed to converge in MAXFN iterations (signifies difficulty in eigenvalue estimation)
4 In ZBRENT, f(a) and f(b) have the same sign (signifies difficulty in eigenvalue estimation)
5 Negative pivot encountered in factorization


next up previous contents
Next: Notes on Use Up: NSPCG User's Guide Previous: Using Multicolor Orderings