Skip to content

Question about baseline overhead costs of sable #1477

Description

@elmerfuddbillwilkins

Currently running Neoforge 1.21.1, getting pretty poor server performance.

In a fresh world with no sublevels, a spark profiler for ticks above 250ms, Sable and sable_rapier are taking 23 and 21% of server thread respectively, though it is mostly through the shared Rapier3D.step() method at 18% thread.

Even with a more general >100ms profiler, sable is still at 9% thread even though no sublevels exist

Im wondering if this is normal behaviour I should accept or something is awry

https://spark.lucko.me/5M9IaPayV8
https://spark.lucko.me/oJ7t5M3TRh

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