Hello, 

During the time elapsed I have got some more clues. In my program, I am using the arma estimations. And this problem occurs in the case when my dependent variable is constant for a long time. When I investigate my log file it says: 

Kalman bla bla returned 36

Deleting var 8 (x1) at list pos 8: all zero 


The weird thing is I am handling this situation (If arma returns some  number other than 0, i am making persistent estimation). When I look at the results, it seems that it is handled. However, When I look at the task

manager, It seems that Gretl is still working and consuming 98% of CPU. As I said before, This situation was not present on 32 bit Ubuntu. 


I am very thankful for your helps. 


2012/4/3 Allin Cottrell <cottrell@wfu.edu>
On Tue, 3 Apr 2012, Emrah Samdan wrote:

> I have developed a code using the libgretl and it was running on a Ubuntu
> 11.04 32 bit machine. Then I migrated those code to Fedora 16(64 bit) and
> compiled it in that machine also. However, while it was running on Fedora,
> sometimes it doesn't quit the program as expected and utilizes 98% of CPU.
> I couldn't get the reason behind this. There was no problem while running
> on 32 bit Ubuntu.
>
> Do you have any idea about this situation ?

Hmm, I'm running current CVS gretl on 64-bit fedora 16 and I'm
not seeing any problems.

When I started using gretl on a 64-bit system I noticed one
"64/32" bug: the mechanism for importing data from a RATS
database was broken due to use of the "long int" type where a
32-bit integer was wanted. But that was fixed in the gretl
1.9.8 release and I'm not aware of any other word-size issues.

Allin Cottrell

_______________________________________________
Gretl-devel mailing list
Gretl-devel@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-devel



--
---
Emrah Samdan