RE: Questions about the raw output file in NM7
I would recommend the following:
1. Set PRINT=9999
This way, iteration 0 is printed, as well as the final results.
Unlikely you will have more than 9999 iterations to have any intervening
iterations printed.
2. Do not specify the FILE= in a multi-problem statement such as the
$SIML NSUB=3 type. The .ext file name will be defaulted to the root
name of your control stream file, so you do not have control of its
name, but it should not be over-written from one problem to the next, or
in a multi-problem statement.
Robert J. Bauer, Ph.D.
Vice President, Pharmacometrics
ICON Development Solutions
Tel: (215) 616-6428
Mob: (925) 286-0769
Email: [email protected]
Web: www.icondevsolutions.com
Quoted reply history
-----Original Message-----
From: [email protected] [mailto:[email protected]]
On Behalf Of Sebastien Bihorel
Sent: Wednesday, March 10, 2010 4:28 PM
To: [email protected]
Subject: [NMusers] Questions about the raw output file in NM7
Dear NMusers,
I would like to have your help on two issues regarding the raw output
file in NM7:
1- when using $ESTIMATION PRINT=0 with "old" estimation methods (FO,
FOCE, FOCEI, HYBRID and Laplacian), the raw output files only contain
the -1000000000 line plus possibly the -1000000001, -1000000002 and
-1000000003 lines. With new estimation methods, the files also contain
data at iteration 0. Is there an option we can use with old estimation
method to request the iteration 0 data to be written into the raw output
files?
2- when an estimation step is requested after a simulation step with
NSUB>1, the raw output file only reports the result of the last
simulation/estimation run. E.g. when I tried the following code,
$SIM (1212345) NSUB=3
$ESTIMATION METHOD=COND FILE=whatever.est PRINT=1 MAXEVAL=5000 NOABORT
only TABLE NO. 3 was reported in whatever.est. By stopping the run at
different times, I found out that the content of the file is actually
overwritten at each sub-problem. Is there a option that would append the
new content to the file rather than overwriting it?
Thank you in advance for your feedback.
Sebastien Bihorel