Table of Contents

Delegate PolyChebNodesKernel

Namespace
NumSharp.Backends.Kernels
Assembly
NumSharp.dll

One Chebyshev node fill (see NumSharp.Backends.Kernels.PolyChebNodes for each formula): count float64 results written contiguously from dst, element k computed from the exact ramp value first + k*step (or k itself, for NumSharp.Backends.Kernels.PolyChebNodes.Pts2).

public delegate void PolyChebNodesKernel(double* dst, long count, double first, double step, double scale)

Parameters

dst double*

Element 0 of count contiguous float64 slots, every one written.

count long

Element count (zero or less writes nothing).

first double

The ramp's first value (NumSharp.Backends.Kernels.PolyChebNodes.Pts1, NumSharp.Backends.Kernels.PolyChebNodes.Gauss) or the added start (NumSharp.Backends.Kernels.PolyChebNodes.Pts2).

step double

The ramp's step, or linspace's step for NumSharp.Backends.Kernels.PolyChebNodes.Pts2.

scale double

The multiplier (NumSharp.Backends.Kernels.PolyChebNodes.Pts1) or divisor (NumSharp.Backends.Kernels.PolyChebNodes.Gauss).

Remarks

The running ramp value is advanced by addition, which stays exact while every value is an integer below 2^53 in magnitude — true for every count an allocation can hold (2^53 doubles are 64 PiB). A caller passing a non-integral first or step would get the rounding of repeated addition, not of first + k*step: this kernel is for NumPy's integer ramps only.