Dear all,
I am conducting a scm search of covariates using PsN. The model uses ADVAN6
with FO estimation method. Due to the complexity of the model, it crashes
after 13 rounds of running with an error message saying:
out of memory!
The scm uses a parallel processing of 5 CPU cores. At the beginning of the
model we stated $SIZE as below:
$SIZES LIM6=3224 LTH=70 LVR=20 NO=1000
The base model has 15 thetas and 8 ETAs. After 13 rounds of scm, the model
has 38 thetas, which is well below LTH. After checking the last run, all
the NONMEM RUNS were finished in the m1 folder. I was wondering if anybody
has encountered such a problem before? Is this because of inappropriate
resource allocation, or because of the memory shortage? Actually, I am not
even sure whether this is because of error in PsN, or error in the NONMEM
itself.
Some other details of the model:
1. scm file investigates effect of 5 covariates on 8 parameters (5*8);
2. NONMEM is ver 7.2.
Many thanks for your kind help.
Best Regards
--
Xinting
PsN Stop: Out of memory.
Dear Xinting,
The "out of memory" error is a Perl error which is caused by the PsN
scm program. Following the [included_relations] procedure described
in my previous message will solve the problem.
Best regards,
Kajsa
On 05/19/2014 05:24 PM, Xinting Wang
wrote:
Dear
all,
I
am conducting a scm search of covariates using PsN. The model
uses ADVAN6 with FO estimation method. Due to the complexity
of the model, it crashes after 13 rounds of running with an
error message saying:
out
of memory!
The
scm uses a parallel processing of 5 CPU cores. At the
beginning of the model we stated $SIZE as below:
$SIZES
LIM6=3224 LTH=70 LVR=20 NO=1000
The
base model has 15 thetas and 8 ETAs. After 13 rounds of scm,
the model has 38 thetas, which is well below LTH. After
checking the last run, all the NONMEM RUNS were finished in
the m1 folder. I was wondering if anybody has encountered such
a problem before? Is this because of inappropriate resource
allocation, or because of the memory shortage? Actually, I am
not even sure whether this is because of error in PsN, or
error in the NONMEM itself.
Some
other details of the model:
1.
scm file investigates effect of 5 covariates on 8 parameters
(5*8);
2.
NONMEM is ver 7.2.
Many
thanks for your kind help.
Best Regards
--
Xinting
--
-----------------------------------------------------------------
Kajsa Harling, PhD
System Developer
Department of Pharmaceutical Biosciences
Uppsala University
[email protected]
+46-(0)18-471 4308
http://www.farmbio.uu.se/research/researchgroups/pharmacometrics/