Skip to content

exp and log function accuracy is too low for double-double and quad-double #438

@Ravenwater

Description

@Ravenwater

double-double should yield 32 digits of precision and quad-double should yield 64 digits of precision. Current function approximation precision is around 14 digits.

The exp() function is at the base of log and pow functions as well, so fix that one first.

Metadata

Metadata

Labels

enhancementnumerical accuracyfunctional issues regarding numerical accuracynumerical investigationInvestigation to explain/quantify numerical properties of a calculation

Type

No type

Projects

Status

Ready

Milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions