Skip to content

Peform extra datatype optimizations for SSA(2) #330

Open
@jasoncarr0

Description

@jasoncarr0

There are several opportunities for late datatype improvements that are not being taken advantage of:

  • Eliminating redundant fields, potentially deeply nested, in datatypes
  • Turning tuples of variants into variants of tuples, when only constructed for a subset of possibilities
  • Unflattening elements which are only used to build another datastructure. (For instance, fields x y and z which are used only to place them in another structure, should likely be folded out into a tuple to save space and time).

This would be partially enabled by eliminating tuples from the IR: #329

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions