On Tue, 3 Jun 2014, Henrique Andrade wrote:
Dear Gretl Community,
I would like to know if the following "strategies" could improve
my scripts in terms of the execution time:
The general answer is, please experiment and tell us what you find.
But I'll venture a few comments.
(1) Write/open gdt files is faster than csv files?
Maybe. The big potential speed-up would be using the gretl gdtb
(binary) format, but this applies only for big datasets.
(2) Using Gretl's home directory (located at my hard drive) is
faster than
use a network drive?
Depends on hardware and software of which I have no knowledge.
Maybe.
(3) Specify the object type could reduce the execution time? (e.g.
"scalar
x = 2" instead of "x = 2")
In principle yes, I think that should give a speed-up. Whether it's
noticeable, I don't know.
(4) Short variables/lists names could reduce the execution time?
I doubt whether that would be noticeable; but it would be
interesting to know.
(5) Matrices are longer than series (I'm talking about computer
memory
usage)?
Not sure what you mean there. Both gretl matrices and gretl series
are pretty "light" in memory usage; neither of them require much
more than (number of numerical values) times (size of
double-precision floating point value).
Allin Cottrell