I run both ADMB version 9 and 10 on MinGW compiler on windows, your example in list (2) can not create .std file on both version and both complain. <div>also your examples in list(5) both worked on the two version, which they should. So I guess it may due to the different compiler, you are using borland. The other thing is you may have to delete the .std file before you run the test, or at least check the time stamp of the file.</div>
<div><br></div><div>weihai <div><br></div><div><br><br><div class="gmail_quote">On Mon, Nov 7, 2011 at 9:39 AM, Larry Jacobson <span dir="ltr"><<a href="mailto:ljacobso@mercury.wh.whoi.edu">ljacobso@mercury.wh.whoi.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Folks-<br>
<br>
1) I have an observation, a suggestion and a request for information about convergence and std file generation in ADMB Version 10 in Windows (Borland compiler).<br>
<br>
2) Observation: Version 10 in Windows does not give an std file after in some cases were Version 9 did.<br>
<br>
In particular, Version 10 seems to be pickier and will not generate an std file unless all of the parameters and sdreport objects are fully converged. For example, code like:<br>
<br>
PARAMETER_SECTION<br>
sdreport_number xx<br>
PROCEDURE_SECTION<br>
xx=0.0;<br>
<br>
would generate an std file under version 9 (with the complaint "Estimated covariance matrix may not be positive definite") but version 10 does not (and it doesn't complain).<br>
<br>
3) Suggestion: I don't have a problem with the new approach but it would be helpful to get a diagnostic or a message. It would be particularly good if the offending parameter(s) were identified (is this possible?).<br>
<br>
4) Request: Can someone describe the current algorithm for determining if the std file is written? Is there anything about the algorithm that might help locate problems quickly?<br>
<br>
5) I attach a real world example (from Mike Wilberg at the University of Maryland, Chesapeake Bay Laboratory). This code sets certain elements of an sdreport_vector called qr to zero (which admittedly it shouldn't do). Anyway, the tpl prefixed NO_ generates an std file under version 9 but not under version 10. The tpl prefixed YES_ generates an std file under both versions. The only difference between the two is that the vector qr is a vector in NO_ and an sdreport_vector in YES_ (see line 167).<br>
<br>
6) I am getting anal in my old age...<br>
<br>
Cheers!<br>
<br>
<br>
-- <br>
<br>
<br>
**********************<br>
"The contents of this message do not reflect<br>
any position of the Government or NOAA."<br>
**********************<br>
Larry Jacobson<br>
National Marine Fisheries Service<br>
Northeast Fisheries Science Center<br>
166 Water Street<br>
Woods Hole, MA 02543-1026<br>
Voice: <a href="tel:508-495-2317" value="+15084952317" target="_blank">508-495-2317</a><br>
Fax: <a href="tel:508-495-2393" value="+15084952393" target="_blank">508-495-2393</a><br>
E-mail: <a href="mailto:larry.jacobson@noaa.gov" target="_blank">larry.jacobson@noaa.gov</a><br>
**********************<br>
<br>
<br>_______________________________________________<br>
Users mailing list<br>
<a href="mailto:Users@admb-project.org">Users@admb-project.org</a><br>
<a href="http://lists.admb-project.org/mailman/listinfo/users" target="_blank">http://lists.admb-project.org/mailman/listinfo/users</a><br>
<br></blockquote></div><br></div></div>