127 commits

Author	SHA1	Message	Date
fredrik	7b31f65f88	add acb_agm	2020-03-13 13:32:20 +01:00
fredrik	b17506e22d	add rel_one_accuracy_bits methods; improvement to precision handling for hypgeom_m	2019-02-22 13:28:03 +01:00
fredrik	242fa2e4c0	header file cleanup (de-inlining, etc)	2019-01-18 15:34:54 +01:00
fredrik	ea2c300a1f	fix fallback code for sub_dddmmmsss	2018-09-17 11:33:49 +09:00
fredrik	7c98883478	approximate dot product and matrix multiplication	2018-09-07 18:04:30 +02:00
fredrik	fb9a514763	32-bit bugfix for acb_dot	2018-09-06 15:46:48 +02:00
fredrik	ffa60bf319	implement acb_dot and use in the acb_poly module	2018-08-26 19:59:30 +02:00
fredrik	85abb2ba84	move mag_add_ui_lower; small cleanup	2018-04-10 17:49:06 +02:00
fredrik	e074eeaec6	acb_rsqrt: more accurate algorithm for wide intervals	2018-03-25 19:59:59 +02:00
fredrik	97c263a7cc	small code cleanup	2018-03-05 13:36:33 +01:00
Ralf Stephan	81a0bcca31	use acb_neg	2018-03-05 07:14:47 +01:00
Ralf Stephan	0673fc935f	fix acos	2018-03-04 17:46:31 +01:00
Ralf Stephan	6eabee7efa	add missing clear	2018-03-04 10:15:52 +01:00
Ralf Stephan	ab6647e29b	rearrange code	2018-03-04 10:00:18 +01:00
Ralf Stephan	1bd2179a46	return exact zero where possible: acos, acosh	2018-03-04 09:15:50 +01:00
fredrik	63526714e2	utility functions acb_sqrt_analytic, acb_rsqrt_analytic, acb_log_analytic, acb_pow_analytic, acb_real_sqrtpos; also improve bounds in acb_rsqrt	2018-02-23 00:21:01 +01:00
fredrik	73a8d86da8	add arb_log_hypot and improve acb_log for wide intervals	2018-02-21 19:03:26 +01:00
fredrik	e11d502161	slight improvement for wide intervals in gamma function	2018-02-21 14:08:54 +01:00
fredrik	dc57bf64a2	fix acb_real_floor / acb_real_ceil (for some reason the wrong code was committed)	2018-02-13 10:42:44 +01:00
fredrik	760d30c171	small optimization of acb trig functions for pure real or imaginary input	2018-02-13 10:11:21 +01:00
fredrik	fbe994f446	utility functions acb_real_abs, sgn, heaviside, floor, ceil, min, max	2018-02-10 22:31:33 +01:00
Fredrik Johansson	1535e13956	faster implementation of acb_get_mag_lower, plus more test code	2017-12-13 11:37:05 +01:00
Fredrik Johansson	6a9ba9d5f4	arb_pow and acb_pow: invert before powering with binary exponentiation to reduce blowup	2017-11-27 16:09:40 +01:00
Fredrik Johansson	a0ce21c37c	more trig functions (sec, csc, sech, csch)	2017-11-13 01:59:40 +01:00
Pascal	226ed94603	order arguments in acb_vec_unit_roots	2017-10-30 08:43:01 +01:00
Pascal	4376829974	change acb_vec_unit_roots interface ... separate order from length, and allow negative order	2017-10-30 08:43:00 +01:00
Fredrik Johansson	aa30e792ae	acb_atan: improve error bounds	2017-09-18 00:06:26 +02:00
Fredrik Johansson	ac713b31aa	speed up acb_agm1 by finishing with Taylor series	2017-08-12 21:05:54 +02:00
Fredrik Johansson	683dea6c17	agm: use central difference instead of forward difference to reduce precision	2017-08-08 23:36:56 +02:00
Fredrik Johansson	de0349a885	add arb_sinc_pi, acb_sinc_pi	2017-06-20 09:44:25 +02:00
Fredrik Johansson	af1f607c32	add acb_dirichlet_hurwitz; rearrange internal call sequence to make sure Riemann zeta code is used in all cases when a=1	2017-06-19 10:23:44 +02:00
Fredrik Johansson	e08d863d58	add acb_expm1	2017-06-12 16:40:33 +02:00
Fredrik Johansson	919e6179a5	some further lambertw precision tweaks	2017-04-22 13:59:37 +02:00
Fredrik Johansson	8dcf5b383a	acb_lambertw: avoid some redundant computations	2017-03-30 21:54:43 +02:00
Fredrik Johansson	fd3801531c	acb_lambertw: slightly better check for using -1/e series	2017-03-30 09:10:48 +02:00
Fredrik Johansson	12adc8747d	acb_lambertw: slightly better derivative bound for |k| <= 1	2017-03-30 09:05:57 +02:00
Fredrik Johansson	4528967782	support nonstandard branches in acb_lambertw	2017-03-28 01:44:41 +02:00
Fredrik Johansson	0fd1e0b833	some improvements to acb_lambertw; make arb_nonnegative_part public	2017-03-23 00:44:49 +01:00
Fredrik Johansson	d77b6c5c73	complex Lambert W function (acb_lambertw)	2017-03-20 18:57:53 +01:00
Tommy Hofmann	6bf072eb59	Replace abort with flint_abort. ... This will allow us to not loose the julia session on error. See also https://github.com/wbhart/flint2/pull/243	2017-02-28 16:52:57 +01:00
Fredrik Johansson	0be7de5ec4	de-inline some more arb and acb methods	2017-02-26 18:18:56 +01:00
Fredrik Johansson	f0570449f9	acb_inv: tighter error bounds (in particular, ensure nonzero input -> finite output)	2017-02-18 17:27:17 +01:00
Fredrik Johansson	249830451f	acb_sqrt: use more accurate formula in the left half plane	2017-02-15 16:36:32 +01:00
Fredrik Johansson	8f9f343127	add arb_is_int_2exp_si and acb_is_int_2exp_si	2017-02-14 14:31:48 +01:00
Fredrik Johansson	8f6fa5598c	acb_nth_root -> acb_unit_root (clearer since there is also a function for computing the nth root of an arbitrary number)	2016-12-01 21:53:02 +01:00
Fredrik Johansson	dfc7965898	_acb_vec_nth_roots: simpler, better algorithm	2016-11-20 05:09:02 +01:00
Fredrik Johansson	06df4d6dcb	wrapper for zeta algorithms	2016-10-25 21:57:37 +02:00
Isuru Fernando	36357d6abd	Merge remote-tracking branch 'upstream/master' into cmake2	2016-10-14 17:47:20 +05:30
Pascal	0fd358048f	move dft to separate directory	2016-10-08 22:49:40 +02:00
Fredrik Johansson	230ec151c6	acb_printn: skip parts that are 0	2016-09-20 14:15:30 +02:00