Skip to content

More efficient memory access for DiskBacking #61

@viking-sudo-rm

Description

@viking-sudo-rm

Right now, whenever we read/write a field on a node, we read/write the whole thing rather than just part of it.

However, a lot of the code just accesses individual fields (like the failure transition) at a time. It would be nice to change it so that we only read/write the field we need. There are maybe two ways to do this:

  1. Only read specific byte indices from a Node or Edge struct stored in a vector of nodes and edges
  2. Separate out the different attributes of nodes and edges into different files (lengths.vec, failures.vec, etc.) and accessing elements of those.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions