This is the mail archive of the gsl-discuss@sources.redhat.com mailing list for the GSL project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: GSL 1.4: BUG #8 [rk8pd.c]




M Joonas Pihlaja wrote:
On Sat, 8 Nov 2003, Luc Maisonobe wrote:

[snip]

Despite I think nobody is interested anymore, I promised to check the
coefficients of the method, so here are my results.


I'm interested, at least, so thank you for sharing.

I am still interested because I currently using it: the rk4 series did not pass my test whereas the rk8pd passed it. Note that my test is rather specific to what I am doing: the test just checks the conservstion of one quatity (namely the energy). Nevertheless I was confused by the BUG report, hence my email to gsl-discuss. So any precision is welcome.



The maximal error in rk8pd.c is about 3.92e-17.
The maximal error in rksuite.f is about 4.49e-29.


What do you mean by maximal error here -- of the main estimator
itself or the second embeded estimator?  And how do you compute
these -- from the principal error function or some other method?
I'm very curious, because you say the results for the
coefficients are exact after fixing two free parameters.

Best Regards,

Joonas Pihlaja


Thanks, Jerome BENOIT


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]