Skip to content

Use full double precision in MFILE.DAT output #3129

@jonmaddock

Description

@jonmaddock

Currently, MFILE.DATs only contain values to 5 significant figures. This is a substantial loss of the precision afforded by double floats, which have a machine epsilon $\epsilon_m \approx 2.22e-16$. This means that 17 significant figures are required to fully represent the precision of the doubles used in Process.

The OUT.DAT is human-readable so can be formatted; the MFILE.DAT should contain full precision. This is necessary to ensure a precise definition of the solution optimisation parameter vector in the MFILE, which is important for reliability analysis studies.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions