lacaml / TODO.md

  • Improve performance with threads on small data

Currently all external functions that have at least linear running time wrt. some parameter release the OCaml runtime lock. This coarse-grained approach may make Lacaml perform less well with some functions on small amounts of data, and also in applications that link with threads, but do not really use them at runtime. This problem is especially pronounced if the user has to deal with large numbers of small datasets and can lead to quite noticable slowdowns then.

Solution: allow the user to explicitly request that the OCaml runtime lock be kept, and also provide for default heuristics that choose to keep it if some value computed from the function parameters does not exceed a certain value (e.g. matrix size, etc.).

Tip: Filter by directory path e.g. /media app.js to search for public/media/app.js.
Tip: Use camelCasing e.g. ProjME to search for ProjectModifiedEvent.java.
Tip: Filter by extension type e.g. /repo .js to search for all .js files in the /repo directory.
Tip: Separate your search with spaces e.g. /ssh pom.xml to search for src/ssh/pom.xml.
Tip: Use ↑ and ↓ arrow keys to navigate and return to view the file.
Tip: You can also navigate files with Ctrl+j (next) and Ctrl+k (previous) and view the file with Ctrl+o.
Tip: You can also navigate files with Alt+j (next) and Alt+k (previous) and view the file with Alt+o.