headers
Patrick Alken | 23 May 2013 22:32
Picon

[bug #39057] gsl_cdf_chisq_Pinv fails for some values

URL:
  <http://savannah.gnu.org/bugs/?39057>

                 Summary: gsl_cdf_chisq_Pinv fails for some values
                 Project: GNU Scientific Library
            Submitted by: psa
            Submitted on: Thu 23 May 2013 08:32:36 PM GMT
                Category: Runtime error
                Severity: 3 - Normal
        Operating System: all
                  Status: None
             Assigned to: None
             Open/Closed: Open
                 Release: 
         Discussion Lock: Any

    _______________________________________________________

Details:

The following parameters cause gsl_cdf_chisq_Pinv to fail - this has been
commented out of cdf/test.c to enable 'make check' to succeed.

  /* Test case reported by Yan Zhou <zhouyan <at> me.com> */
  TEST (gsl_cdf_chisq_Pinv, (0.5, 0.01), 0.99477710813146, TEST_TOL6);

    _______________________________________________________

Reply to this item at:
(Continue reading)

Permalink | Reply | 0 comments |
headers
Patrick Alken | 23 May 2013 22:26
Picon

[bug #39056] gsl_sf_hyperg_2F1_e fails for some test cases

URL:
  <http://savannah.gnu.org/bugs/?39056>

                 Summary: gsl_sf_hyperg_2F1_e fails for some test cases
                 Project: GNU Scientific Library
            Submitted by: psa
            Submitted on: Thu 23 May 2013 08:26:48 PM GMT
                Category: Runtime error
                Severity: 4 - Important
        Operating System: all
                  Status: Confirmed
             Assigned to: None
             Open/Closed: Open
                 Release: 
         Discussion Lock: Any

    _______________________________________________________

Details:

The following test cases are confirmed to fail for the gsl_sf_hyperg_2F1_e
function. They have been commented out of specfunc/test.c to enable 'make
check' to succeed.

  /* Test case from Hatef Monajemi <monajemi <at> stanford.edu> */

  TEST_SF(s, gsl_sf_hyperg_2F1_e, (3.5, -0.5, 5.0, 0.9, &r),
0.5923981284370653465208973272, TEST_TOL2, GSL_SUCCESS);

  /* Test case from Robert L Wolpert <Wolpert <at> stat.duke.edu> */
(Continue reading)

Permalink | Reply | 0 comments |
headers
Patrick Alken | 23 May 2013 22:14
Picon

[bug #39055] gsl_poly_complex_solve fails on a 15th degree polynomial

URL:
  <http://savannah.gnu.org/bugs/?39055>

                 Summary: gsl_poly_complex_solve fails on a 15th degree
polynomial
                 Project: GNU Scientific Library
            Submitted by: psa
            Submitted on: Thu 23 May 2013 08:14:21 PM GMT
                Category: Runtime error
                Severity: 3 - Normal
        Operating System: all
                  Status: Confirmed
             Assigned to: None
             Open/Closed: Open
                 Release: 
         Discussion Lock: Any

    _______________________________________________________

Details:

The following code demonstrates a failure of gsl_poly_complex_solve(). I have
commented this out from poly/test.c since it will cause that test to fail
until the problem is fixed.

The QR reduction of the polynomial companion matrix does not converge. The
eigenvalues of the companion matrix can be successfully computed with
gsl_eigen_nonsymm(), but this is not a good long-term solution - the qr
algorithm should be fixed somehow. Until it is I think this bug report is best
left here rather than in poly/test.c where it causes 'make check' to fail.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Rudy Arthur | 19 May 2013 22:49
Picon

MISER subdivision bug

Hi,

I found what I think is a bug in miser.c.
In the function estimate_corrmc, line 689 in the file
  sigma_l[i] = sqrt (fsum2_l[i] - fsum_l[i] * fsum_l[i] / hits_l[i]);
is supposed to be the variance in the left hand region with the division in
the i-axis ,
Shouldn't this be:
  sqrt( ( fsum2_l[i] / hits_l[i] - fsum_l[i] * fsum_l[i] ) / (hits_l[i] -
1) )
and correspondingly for the right hand variance? Or is there some subtlety
I am missing?

Rudy.
Permalink | Reply | 0 comments |
headers
Michael Bunk | 30 Apr 2013 14:13
Picon
Favicon

ATOM-Feed does not validate

Hello,

please check
http://validator.w3.org/feed/check.cgi?url=http%3A%2F%2Fsavannah.gnu.org%2Fnews%2Fatom.php%3Fgroup%3Dgsl

as it says:

Sorry

This feed does not validate.

    line 228, column 100: XML parsing error: <unknown>:228:100:
undefined entity [help]

        ... UINT_MAX [<em><a href="/bugs/?24897">bug&nbsp;#24897</a></em>]

-> HTML entities are not allowed in plain XML...

Best regards,
 Michael Bunk
Attachment (smime.p7s): application/pkcs7-signature, 4064 bytes
Permalink | Reply | 0 comments |
headers
David Faustner | 30 Apr 2013 14:03
Picon
Favicon

Question concerning code sequences in GSL

Dear GNU Scientific Library Team,

I've been studying section 35.3: "Providing the function to solve"  in the
GNU Scientific Library Reference Manual and I deal with a command sequence,
where I'm not sure if it is stated correctly.  My question relates to the
powell function according to 

     struct powell_params { double A; };

     int

     powell (gsl_vector * x, void * p, gsl_vector * f) {

        struct powell_params * params

          = *(struct powell_params *)p;

        const double A = (params->A);

        const double x0 = gsl_vector_get(x,0);

        const double x1 = gsl_vector_get(x,1);

        gsl_vector_set (f, 0, A * x0 * x1 - 1);

        gsl_vector_set (f, 1, (exp(-x0) + exp(-x1)

                               - (1.0 + 1.0/A)));

        return GSL_SUCCESS
(Continue reading)

Permalink | Reply | 1 comment |
headers
eric | 30 Apr 2013 13:02
Picon

bug report

Hi there!
I got the following failure after running 'make check > log 2>&1'.
My Operating System is Debian 6.0 amd64.

Making check in sort
make[1]: Entering directory `/home/eric/Downloads/gsl-1.9/sort'
make  test
make[2]: Entering directory `/home/eric/Downloads/gsl-1.9/sort'
gcc -DHAVE_CONFIG_H -I. -I. -I.. -I.. -I..    -g -O2 -c test.c
/bin/bash ../libtool --tag=CC --mode=link gcc  -g -O2   -o test  test.o
libgslsort.la ../permutation/libgslpermutation.la ../vector/libgslvector.la
../block/libgslblock.la ../ieee-utils/libgslieeeutils.la ../err/libgslerr.la
../test/libgsltest.la ../sys/libgslsys.la -lm 
gcc -g -O2 -o test
test.o  ./.libs/libgslsort.a ../permutation/.libs/libgslpermutation.a
../vector/.libs/libgslvector.a ../block/.libs/libgslblock.a
../ieee-utils/.libs/libgslieeeutils.a ../err/.libs/libgslerr.a ../test/.libs/libgsltest.a
../sys/.libs/libgslsys.a -lm
make[2]: Leaving directory `/home/eric/Downloads/gsl-1.9/sort'
make  check-TESTS
make[2]: Entering directory `/home/eric/Downloads/gsl-1.9/sort'
FAIL: indexing gsl_vector_char, n = 128, stride = 1, ordered [19999]
FAIL: sorting, gsl_vector_char, n = 128, stride = 1, ordered [20000]
FAIL: smallest, gsl_vector_char, n = 128, stride = 1, ordered [20001]
FAIL: largest, gsl_vector_char, n = 128, stride = 1, ordered [20002]
FAIL: smallest index, gsl_vector_char, n = 128, stride = 1, ordered
[20003]
FAIL: largest index, gsl_vector_char, n = 128, stride = 1, ordered
[20004]
FAIL: indexing gsl_vector_char, n = 128, stride = 1, reversed [20005]
(Continue reading)

Permalink | Reply | 1 comment |
headers
Ming Ji | 24 Apr 2013 18:23
Picon

undesired behavior of gsl_stats_sd

Hi,

The output of the following test code is nan. I don't think this is the
desired behavior for users. When number of input data is 1, I think it is
more reasonable to say standard deviation is 0 instead of nan.

thanks,
Ming

#include <iostream>
#include <gsl/gsl_statistics.h>

itn main()
{
    double data[1] = {1.0};
    double sd = gsl_stats_sd(data, 1, 1);
    std::cout << sd << std::endl;
    return 0;
}
Permalink | Reply | 2 comments |
headers
Pongetti, Thomas J (3286 | 23 Apr 2013 21:40
Picon
Picon
Favicon

(no subject)

make  check-TESTS
make[2]: Entering directory `/clars/gsl-1.15/histogram'
FAIL: gsl_histogram_fprintf and fscanf [36]
FAIL: test.exe
==================
1 of 1 test failed
==================
Makefile:453: recipe for target `check-TESTS' failed
make[2]: *** [check-TESTS] Error 1
make[2]: Leaving directory `/clars/gsl-1.15/histogram'
Makefile:575: recipe for target `check-am' failed
make[1]: *** [check-am] Error 2
make[1]: Target `check' not remade because of errors.
make[1]: Leaving directory `/clars/gsl-1.15/histogram'
Permalink | Reply | 0 comments |
headers
Martin Wiebusch | 19 Apr 2013 02:43
Picon

wrong formula for BFGS update in gsl_multimin?

I've had some trouble with non-converging minimizations using the
gsl_multimin_fdfminimizer_vector_bfgs minimizer, which finally led me to
look into the code. The comments in 'vector_bfgs.c' and 'vector_bfgs2.c'
read

/* This is the BFGS update: */
/* p' = g1 - A dx - B dg */
/* A = - (1+ dg.dg/dx.dg) B + dg.g/dx.dg */
/* B = dx.g/dx.dg */

where, as far as I can see, p' is the new search direction, g1 and g
both denote the new gradient, dx is the position change and dg the
change in the gradient after the line search.

All the literature I've found on the BFGS method defines the new search
direction as p' = B^(-1) g, where B^(-1) is the BFGS estimate of the
inverse Hessian, defined trough outer and inner products of dx and dg
and the inverse Hessian B_prev^(-1) of the _previous_ iteration. (See,
e.g. http://en.wikipedia.org/wiki/BFGS_method). Using these expressions,
I can only reproduce the formula from the comments for the special case
where B_prev^(-1) is the identity matrix. I've also compared, during an
actual minimization, the vectors B^(-1) g (with my own implementation of
the B^(-1) matrix update) and the vector p' from the formula in the
comments, and they are indeed different.

Is it possible that the BFGS update in GSL is incorrect?

Best regards,
Martin.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Thomas Beutlich | 18 Apr 2013 16:59
Picon
Favicon

Memory optimization in akima interpolation

I checked the akima memory allocation and found some potential optimizations:

 * It is sufficient to allocate (size - 1) doubles for the spline coefficients b, c and d of akima_state_t.

 * It is sufficient to allocate (size + 3) doubles for the divided differences _m of akima_state_t.

 * The divided differences _m of akima_state_t are only accessed in akima_calc and akima_init /
akima_init_periodic, hence could be freed earlier at the end of the calculation of the spline
coefficients b, c and d in akima_calc.

Best regards,
Thomas Beutlich

--

http://tbeu.de
Permalink | Reply | 0 comments |

Gmane