error-analysis
function γ(n: number): number
Defined in error-analysis/error-analysis.ts:40
The canonical floating point error function, γ.
- very close to being
=== n * (Number.EPSILON / 2) - see e.g. Algorithms for Accurate, Validated and Fast Polynomial Evaluation
Parameters:
| Name | Type | Description |
|---|---|---|
n | number | the parameter - typically a small positive integer, e.g. for polynomial evaluation this === 2*d + 1, where d is the degree of the polynomial |
function γγ(n: number): number
Defined in error-analysis/error-analysis.ts:58
The canonical, once compensated (implying double-double precision), floating point error function.
- very close to being
=== n * (Number.EPSILON / 2)**2 - see e.g. Algorithms for Accurate, Validated and Fast Polynomial Evaluation
Parameters:
| Name | Type | Description |
|---|---|---|
n | number | the parameter - typically a small positive integer, e.g. for polynomial evaluation this === 2*d + 1, where d is the degree of the polynomial |