Reading GAUSS documentation, svd1 is using the EISPACK code while lapsvd1
is using the LAPACK code. EISPACK (so svd1) is not supported any longer.
No idea how GAUSS is using LAPACK internally. you might hav a problem in
your calling sequence. The message: "lapsvd1: illegal value in argument"
does not sound good for sure ... I am also seeing : "lapsvdusv" maybe you
want to try that one.
On Tue, 22 Feb 2011, julie langou wrote:
Dear Han,you do not seem to use LAPACK from netlib.
Please send you question to team that provided you your Linear Algebra
On Feb 21, 2011, at 3:12 PM, Chen , Han wrote:
Greetings and Salutations,
I use GAUSS program. I use svd1 first and the program was aborted with
error message: "Not all the singular values can be computed." Then I
replace svd1 with lapsvdcusv. The program keeps running with warning
message "lapsvd1: illegal value in argument". Would you be able to tell
me what this means and what could potentially cause this problem? I am
concerned that because of this error message, what I get out of
optimization is garbage. Or could I safely ignore this warning and trust
It is unlikely due to coding error because I cross checked with Dynare
and they gave the same likelihood value at the initial guess.
I really appreciate for your help.
4th year PhD economics student
University of Pennsylvania
Cell: ?+1 (240)-374-2758
160 McNeil, 3718 Locust Walk,
Philadelphia, PA 19104
Lapack mailing list
Julie Langou; Research Associate in Computer Science
Innovative Computing Laboratory;
University of Tennessee from Denver, Colorado ;-)