vquadpy - Vectorized parallel numeric integrator and interpolator

	Commit message (Collapse)	Author	Age	Files	Lines
*	Lyness-Kaganove benchmark	Christoph Groth	2018-12-20	3	-2/+349
\|
*	calculate proper error estimate already when initializing integrator	Christoph Groth	2018-12-14	1	-6/+8
\| \| \| \|	It is no longer necessary to call `improve()` first.
*	_Interval: replace the method points() by map()	Christoph Groth	2018-05-16	1	-5/+6
\|
*	sharpen divergence detection test	Christoph Groth	2018-05-16	1	-26/+28
\| \| \| \| \| \|	Instead of testing the divergence detection with functions that consist of a single peak, use functions of multiple peaks with the exponent drawn randomly.
*	add "battery" test	Christoph Groth	2018-03-03	1	-47/+160
\|
*	add absolute error condition	Christoph Groth	2018-03-03	1	-6/+11
\|
*	turn split and refine into coroutines and move them to _Interval	Christoph Groth	2018-02-28	1	-49/+62
\|
*	implement evaluation of the interpolated integrand	Christoph Groth	2018-02-28	3	-11/+104
\|
*	move interval quality check into refine()	Christoph Groth	2018-02-23	1	-13/+18
\|
*	get rid of nr_points	Christoph Groth	2018-02-23	2	-20/+10
\|
*	use insort instead of sort	Christoph Groth	2018-02-23	1	-2/+8
\|
*	sort ivals instead of using i_max	Christoph Groth	2018-02-23	1	-23/+17
\|
*	no longer delete any intervals	Christoph Groth	2018-02-23	1	-17/+4
\| \| \| \| \| \| \| \| \| \|	The maximum number of intervals is no longer limited. Arbitrarily deleting intervals introduced arbitrary inaccuracies and there's really no point. In typical cases the max_ivals=200 was not reached anyway and in pathological cases killing intervals is not a useful thing to do. When intervals become too narrow for further refinement they are no longer deleted but end up in self.attic.
*	initial commit	Christoph Groth	2018-02-23	9	-0/+655