I have noticed a few anomalies in a few special funtions of GSL: 1) Both gsl_sf_zeta_int() and gsl_sf_zetam1_int() return zero at even negative integers. gsl_sf_zetam1_int_e() should be modified.
2) gsl_sf_laguerre_n(10000001, 2.5, 2.5) returns a wrong answer; -1.5e6 insted of 1.5e6. laguerre_large_n() seems to return the correct answer only when n is even. Multiply by -1 when n is odd? I don't know. 3) gsl_sf_hyperg_1F1(1.2, 1.1e-15, 1.5) returns a wrong answer. hyperg_1F1_renorm_b0() seems incorrect. I have not been able to understand the implementation of Abramowitz+Stegun 13.3.7. 4) hyperg_1F1_large2bm4a() does not give correct answers. The algorithm is implemented correctly, but fails it's task when a in [-1, 1]. It seems to me that a must be large negative for hyperg_1F1_large2bm4a() to produce sensible results. Harald Moseby, Oslo _______________________________________________ Bug-gsl mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-gsl
