Hydra Build Daemon | 2 Feb 2011 13:48
Picon
Picon
Gravatar

Hydra job gnu:gsl-master:build build 881130: Unknown failure

Hi,

This is to let you know that Hydra build 881130 of job gnu:gsl-master:build has changed from 'Success' to
'Unknown failure'.

Complete build information can be found on this page: http://hydra.nixos.org/build/881130

The last 50 lines of the build log are shown at the bottom of this email.

A summary of the build information follows:

Build ID:              | 881130
Nix name:              | gsl-1.14.27
Short description:     | Native Nix build on i686-cygwin
Maintainer(s):         | bug-gsl <at> gnu.org
System:                | i686-cygwin
Derivation store path: | /nix/store/qi69blhgix50wbhmxxhg3fyh6nhh3qy1-gsl-1.14.27.drv
Output store path:     | /nix/store/9i84ly0hwl03a6fklv0nc1gjcr0gq5n9-gsl-1.14.27
Time added:            | 2011-02-02 13:48:58
Build started:         | 2011-02-02 13:48:44
Build finished:        | 2011-02-02 13:48:58
Duration:              | 14s

The build inputs were:

Name        | Type   | Value
------------+--------+-------------------------------------------------------------
hydraConfig | svn    | https://svn.nixos.org/repos/nix/hydra-config/gnu/trunk:25741
nixpkgs     | svn    | https://svn.nixos.org/repos/nix/nixpkgs/trunk:25759
system      | string | i686-cygwin
(Continue reading)

Imad-Eddine Srairi | 4 Feb 2011 17:19
Picon
Favicon

Re: Re: [Help-gsl] Hypergeometric function

Hi,

I could not manage to fix this problem yet, but I have tried to narrow 
down its location. We are here in a special case where conditions of [1] 
are fulfilled i.e.:

a = -0.5 , b =  1.5 , c =  1.0

and : c = 0.5 * a + 0.5 * b + 0.5.

It is easy to check the two results returned by the test case attached 
to this issue against:

M_SQRTPI * gsl_sf_gamma(1.0) / gsl_sf_gamma(0.25) / gsl_sf_gamma(1.25);

to find that the correct answer is the positive one (the one returned 
for (1/2 - epsilon)) and that the error lies in the (1/2 + epsilon) case.

The result returned in this case is produced by call to this function

         return hyperg_2F1_reflect(a, b, c, x, result);

which seems to rely on this reference:

/* Do the reflection described in [Moshier, p. 334].
  * Assumes a,b,c != neg integer.
  */

I have no access to this book, but I am currently trying to match the 
code against the special cases described in [2].
(Continue reading)

Brian Gough | 11 Feb 2011 18:00
Picon

Re: Re: [Help-gsl] Hypergeometric function

At Fri, 04 Feb 2011 17:19:32 +0100,
Imad-Eddine Srairi wrote:
> 
> /* Do the reflection described in [Moshier, p. 334].
>   * Assumes a,b,c != neg integer.
>   */
> 
> I have no access to this book, but I am currently trying to match the 
> code against the special cases described in [2].
> 

Thanks for tracking that down.  I have the book so I will look at it
and see if there is any more information but Moshier doesn't usually
have much explanation of the algorithms.

--

-- 
Brian Gough
Rhys Ulerich | 14 Feb 2011 16:15
Picon

[bug #32237] RFE: Add function for extracting nodes and weights for Gauss-Legendre integration

Update of bug #32237 (project gsl):

                  Status:                    None => In Progress            
             Assigned to:                    None => rhysu                  

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?32237>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/
Rhys Ulerich | 14 Feb 2011 20:25
Picon

[bug #32237] RFE: Add function for extracting nodes and weights for Gauss-Legendre integration

Update of bug #32237 (project gsl):

                  Status:             In Progress => Fixed                  

    _______________________________________________________

Follow-up Comment #1:

I just ran into this need too...

In revision 4711 I implemented gsl_integration_glfixed_point to allow
retrieving Gauss-Legendre points and weights from a
gsl_integration_glfixed_table.  The points are strictly increasing which
should allow you to compare them against other implementations.  Test cases in
integration/test.c check their validity for n = 1, 2, 3, 4, and 5 against some
analytic results as well as ensuring they're scaled correctly for non-[-1,1]
intervals.  Both pass on GCC's -O3.

I'm closing this feature.  Please re-open if you run into any trouble.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?32237>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/
(Continue reading)

Hydra Build Daemon | 24 Feb 2011 23:58
Picon
Picon
Gravatar

Hydra job gnu:gsl-master:build build 937405: Success

Hi,

This is to let you know that Hydra build 937405 of job gnu:gsl-master:build has changed from 'Unknown
failure' to 'Success'.

Complete build information can be found on this page: http://hydra.nixos.org/build/937405

A summary of the build information follows:

Build ID:              | 937405
Nix name:              | gsl-1.14.27
Short description:     | Native Nix build on i686-cygwin
Maintainer(s):         | bug-gsl <at> gnu.org
System:                | i686-cygwin
Derivation store path: | /nix/store/6hrcvxrks7anrsd1jqzixbx5kyg1g06z-gsl-1.14.27.drv
Output store path:     | /nix/store/zhfah20rwm433prrki3vbpmyml8bc2j8-gsl-1.14.27
Time added:            | 2011-02-24 23:58:48
Build started:         | 2011-02-24 23:44:55
Build finished:        | 2011-02-24 23:58:48
Duration:              | 833s

The build inputs were:

Name        | Type   | Value
------------+--------+-------------------------------------------------------------
hydraConfig | svn    | https://svn.nixos.org/repos/nix/hydra-config/gnu/trunk:25975
nixpkgs     | svn    | https://svn.nixos.org/repos/nix/nixpkgs/trunk:26109
system      | string | i686-cygwin
tarball     | build  | 937262

(Continue reading)

Eugeny Abramochkin | 28 Feb 2011 09:58
Picon

Misprint in the GSL Ref.PS

Hi!

Page 34. The definition of the Airy function Bi(x)
The term $e^{-t^3/3+xt}$ instead of  $e^{-t^3/3}$ should be printed in the 
integrand.

Regards.
EGA 
Imad-Eddine Srairi | 2 Mar 2011 23:58
Picon
Favicon

Re: Re: [Help-gsl] Hypergeometric function

Hi,

I looked again at the code of function "hyperg_2F1_reflect" found in 
file hyperg_2F1.c .

As I had noticed and reported in my first e-mail regarding this topic, 
this function handles separately the case where c - a - b = m is an 
integer (which is the case for the bug we are interested in: a = -0.5 ; 
b = 1.5 ; c = 1.0) and computes:

F1 + sgn * F2

where sgn depends on parity of the integer m (sgn = -1 if m is odd, +1 
otherwise).

In the test case we are looking into, F1 = 0.
I am now also pretty sure that F2 is the second part (i.e. the infinite 
sum) of 15.3.11 in Abramowitz&Stegun [1] and of 15.8.10 in the DLMF [2]

There are only two differences. The author of the code has chosen to 
distribute the minus sign: instead of

ln(1-x)-psi(k+1)-psi(k+m+1)+psi(a+k+m)+psi(b+k+m)

(where psi is the digamma function [3]) s/he computes

psi(k+1)+psi(k+m+1)-psi(a+k+m)-psi(b+k+m)-ln(1-x)

I can see nothing wrong with this.

(Continue reading)

Thanh Vo | 3 Mar 2011 19:12
Picon

Floating-point exceptions in the GSL

Hi,

My name's Thanh Vo. My research group is developing a tool to detect
floating point exceptions in numerical code. On Aug 11th 2010, we contacted
the list to report exception triggering inputs we had found in the GSL.
Brian Gough confirmed the exceptions were valid warnings.  We have continued
our work since then.  In this email, we report more, potentially problematic
exception-triggering inputs.

For each exception, we describe the name of function, the type of exception
thrown, the inputs that trigger the exception, and the source line that
throws the exception.

1.  gsl_sf_bessel_Inu_scaled_asymp_unif_e
Invalid: nu = 0.000000e+00 and  x = 0.000000e+00
Location: line 359 at bessel.c

2.  fg_asymp
Divide ByZero: x = 0.000000e+00
Location:  line 228 at sinint.c

3.  gsl_sf_legendre_P3
Overflow: x = -5.996154e+153
Location: line 95 at legendre_poly.c

4.  gsl_sf_gammastar
Underflow: x = 8.988466e+307
Location: line 1342 at gamma.c

You can see the full list of potential bugs at
(Continue reading)


Gmane