Skip to content

Improve memory management through more suitable allocators? #2

Open
@ktnr

Description

@ktnr

https://github.com/r-lyeh-archived/ltalloc seems to work great. Observed runtime improvements in parallel mode with SolveParallelTaskflow() of up to 40%.

I do not yet have a good understanding of the suitability of this allocator, other than it states that it is suited for multi-threaded purposes. Is it similar to pool allocators? Might explicit pool allocators be more useful? Need to study the wiki.

Caution: Need to further evaluate the limitations for usage with Windows. By including the source files the operator new is globally overriden. This shouldn't affect precompiled dependencies. But it might have an effect on all projects that include this repo, and therefore also ltalloc.cc, as a submodule.

Metadata

Metadata

Assignees

No one assigned

    Labels

    performancePerformance improvements

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions