[Psn-general] PsN: optional use of the $EST/MCETA=1 functionality

[Rik S. <rik...@oc...>]

Dear PsN users,

Apologies in advance if this has been discussed at depth previously, but I've just joined the mailing list and couldn't find any reference to this issue… 😊

Using PsN 4.9.0 we just noticed that there seems to be a default speedup functionality implemented by using the MCETA=1 option in $EST, along with a $ETA file likely produced in the initial model file run. When we take individual model runs, they are indeed pretty fast, but they also seem to be pretty wrong. For instance, one of the first covariate runs took only three iterations, but when we rerun the exact same control stream but remove the $ETA line and the MCETA=1 option, it takes 35 iterations, the OFV is 22 points lower and the parameter estimates deviate significantly. And for our particular problem this seems to occur across the board. So the question is: can we turn of this option that clearly was not available in previous PsN versions?

Kind regards,

Rik



Rik Schoemaker, PhD
Occams Coöperatie U.A.
Malandolaan 10
1187 HE Amstelveen
The Netherlands
www.occams.com<http://www.occams.com>
+31 20 441 6410
rik...@oc...<mailto:rik...@oc...>

[cid:image001.png@01D67496.07B04FC0]