macOS security
by Berend Hasselman
H
The current macOS (1.015, Catalina) has additional security checks.
The installer gives a warning for the current gretl-2019d-quartz.pkg distribution.
Apple warns that the program cannot be run. There is however a button to open the file so it can be opened (at your own risk).
But it is not safe to assume that this will continue to be the case.
See this page: https://support.apple.com/en-gb/guide/security/welcome/1/web
As of February 2020 Apple will require that apps distributed outside of the Mac store will have to be notarized.
See https://appleinsider.com/articles/19/12/23/apple-will-enforce-app-notariz...
which most likely means that the current gretl package can't be opened and that gretl cannot be run on macOS Catalina.
Another remark: I do not know how the macOS package is created but the contents cannot be inspected by Pacifist (an app used to view the contents of a ,pkg distribution; see https://charlessoft.com). Very useful app.
regards,
Berend Hasselman
4 years, 7 months
feval and MPI
by Sven Schreiber
Hi,
another MPI question or issue: Is the relatively new 'feval' meta-style
function supported inside MPI? (Of course using the --send-functions
option.)
thanks
sven
4 years, 9 months
speed of (m)ols with svd
by Sven Schreiber
Hi,
I'm observing an unexpectedly drastic slowdown of mols() when switching
to SVD (with 'set svd on').
My toy benchmark is to regress a 200x10 LHS matrix on a 200x80 RHS
matrix 5000 times. The matrices are not redrawn in order to measure just
the mols speed, not the RNG. On an oldish machine I get:
~ 3 sec without svd
~ 60 sec with svd
For comparison, Python/Numpy's linalg.lstsq takes ~ 8 sec and is said to
use SVD always.
There are at least two possible differences between gretl/svd and numpy:
1) numpy might be using Intel's Math Kernel Library (MKL) in the
background instead of standard Lapack (coming from the conda/Anaconda
distro).
2) The web says that numpy uses Lapack's DGELSD
(http://www.netlib.org/lapack/explore-html/d7/d3b/group__double_g_esolve_g...),
whereas judging from lib/src/gretl_matrix.c in the function
gretl_matrix_SVD_ols (and maybe gretl_matrix_multi_SVD_ols? - don't know
what that is for) it seems that gretl uses DGELSS.
On the Netlib Lapack page it says
(https://www.netlib.org/lapack/lug/node27.html) that "The subroutine
xGELSD is significantly faster than its older counterpart xGELSS", and
it labels DGELSD with "solve LLS using divide-and-conquer SVD".
So, any negative side effects blocking the switch to DGELSD, or is it
something like a free lunch?
thanks
sven
4 years, 9 months
MPI reduce syntax documentation
by Sven Schreiber
Hi,
it just took me a while to realize that (inside mpi) I cannot write the
following:
...
string aggrop = "sum"
mpireduce(&m, aggrop)
...
but instead I need to go with string substitution:
mpireduce(&m, @aggrop)
I think the doc on p. 4 of the MPI guide is misleading there when it
says string argument (for mpireduce, mpiallred, mpiscatter). It's rather
a keyword apparently. I'm guessing that a string literal would work, but
haven't tried.
Slightly related about string representations: Consider a string inside
a bundle:
b = defbundle("a", "ohoh")
print b # gives: a = ohoh
The meaning is pretty clear of course, but both the single equal sign
(instead of ":" or perhaps "==") and the non-quotation of this string
literal looks strange to me given the gretl rules.
thanks
sven
4 years, 9 months
problem with sprintf in ternary statement
by Sven Schreiber
Hi,
this is on the Dec 14th snapshot, haven't tested on newer release, sorry.
<hansl>
string s2 = sprintf("\"hey\"")
string snull = 1 ? s2 : "null"
print snull # "hey", as expected
# short form with anonymous sprintf
string snull = 1 ? sprintf("\"hey\"") : "null" # error
print snull
</hansl>
thanks
sven
4 years, 9 months
What to do when convergence is no reached after 800 iterations (nonlinear regression)
by Ivan Falconi
Hello.
Please, I need ssome help.
I am trying to obtain the parameters of the Bass model (p, q , m are
coefficient of innovation, coefficient of imitation and market potential,
respectively) . The message I got was that after 800 iterations convergence
was not reached.
The initial parameters were obtained with Excel's Solver. I had the same
issue before and I used the program NELREG which gave me same error but I
used a script and was able to find convergence. I am sure the same will
happen with greatl provided I use proper script.
I thank you in advance for help.
Regards,
Ivan Falconi
[image: image.png]
4 years, 9 months
next release of gretl
by Allin Cottrell
Hello all,
Some of you will have heard that the 2019d release was planned for
last week. Sorry, but the marking of exams and econometrics projects
has taken its toll. I can guarantee, however, that there will be a
release in time for Christmas! With lots of bug-fixes and quite a
few new features.
Allin
4 years, 9 months
different result printf vs sprintf
by Sven Schreiber
Hi,
here is an extremely unimportant observation, not sure if it is even
supposed to be used in hansl. The same input gives different output in
printf and sprintf:
<hansl>
eval printf("%5s", "a") # gives a5
eval sprintf("%5s", "a") # gives a
</hansl>
thanks
sven
4 years, 9 months
Return of stringified-series includes only integer
by Artur Tarassow
Hi,
(sorry for cross-posting Sven, Jack and Allin -- but I think this got
missed some days ago)
using latest git, I am bite puzzled why the returned series is not
string-values any more.
<hansl>
clear
set verbose off
function series foo (const string s)
strings S = defarray(s)
series ret = 1
stringify(ret, S)
print ret -o # series is string-valued
return ret
end function
nulldata 4
series ser = foo("A")
print ser -o # series includes numerical obs.
</hansl>
Thanks,
Artur
4 years, 9 months
strsplit bug (?)
by Sven Schreiber
Hi,
I'm observing this unexpected behavior:
<hansl>
string test = sprintf("Hey \n ho \n ?")
eval strsplit(test) # 3-elem array, as expected
eval strsplit(test, sprintf("\n")) # only 2 elems!?
</hansl>
This is on the latest snapshot.
thanks
sven
4 years, 9 months
