___________________________________________________________________
| |
| PRCOMW |
|_________________________________________________________________|
MEANING: PREDPP common
CONTEXT: For use with PREDPP
USAGE:
$PK
"FIRST
" COMMON /PRCOMW/ FACTOR
" DOUBLE PRECISION FACTOR
" FACTOR=10.
DISCUSSION:
This feature is not fully documented. The interested user may be able L
to obtain more information by studying the appropriate sections of L
PREDPP code. L
USAGE: L
A message similar to the following is sometimes produced by ADVAN7: L
NUMERICAL DIFFICULTIES OBTAINING THE SOLUTION. NON-REAL L
EIGENVALUES. PERHAPS ADVAN5 SHOULD BE USED FOR THIS PROBLEM. L
"LARGEST" VALUE ALONG SUBDIAGONAL OF SCHUR MATRIX: 0.423E-01 L
(DIAG: 0.84E+00 0.84E+00) L
FACTOR is a "fudge factor" that affects ADVAN7's determination of how L
large the subdiagonal element is relative to the two diagonal ele- L
ments. The default value of FACTOR is 1. L
These error messages arise when the computations of the eigenvalues L
produce unexpected values. The question is whether the subdiagonal L
elements are too large to ignore (e.g., when the eigenvalues are truly L
complex) in which case ADVAN5 must be used, or whether trivially small L
quantities have been produced by the accumulation of rounding error in L
the floating point computation, in which ADVAN7 can be used. (Non- L
real eigenvalues result when the system is non-mamillary, and may also L
occasionally occur when there are etas associated with rate constants L
Kij). L
If the maximum value printed is very small (e.g., E-15), then the com- L
putations will proceed correctly using a value of FACTOR such as 10 or L
100. Values such as FACTOR=10 or FACTOR=100 are more permissive, and L
allow larger subdiagonal elements. If no reasonable value of FACTOR L
eliminates the error messages, then ADVAN5 should be used. L
REFERENCES: None.
Go to main index.
Created by nmhelp2html v. 1.0 written by Niclas Jonsson (Modified by AJB 5/2006,11/2007)