Hello all, To be explicit, I do think that they should be resuable. Each integration routine calls 'initialise()' (the source is located in integration/initialise.c) which should reset the integration workspace for a new integration each time.
One can imagine that the justification for keeping the workspace allocation separate might have been to allow the user the freedom to perform the allocation in their own way, or even to provide a completely different type of integration workspace object. I do think it might be useful to allow the user to specify the subdivisions of the interval before integration to speed things up, but I don't think GSL currently allows this. RADAPT in cernlib does allow this though, and I do occasionally use my C++ version of it at: http://o2scl.sourceforge.net/o2scl/html/classcern__adapt.html Take care, Andrew Steiner On Thu, Dec 1, 2011 at 6:59 AM, Rhys Ulerich <[email protected]> wrote: > Hi Denes, > > On Wed, Nov 30, 2011 at 3:09 PM, Denes Molnar <[email protected]> > wrote: >> The documentation does not discuss whether gsl_integration_workspace is >> reusable. Is it? For example, can one allocate a single workspace, and use >> that to do two different integrations afterwards? Some of our tests with >> gsl-1.14 indicate that the answer is 'no'. > > Do you have some test code demonstrating that 'no' result? > > - Rhys > > _______________________________________________ > Help-gsl mailing list > [email protected] > https://lists.gnu.org/mailman/listinfo/help-gsl _______________________________________________ Help-gsl mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-gsl
