Skip to content

Conversation

@samuelgarcia
Copy link
Member

This try to make lupin working with thread.
This will be usefull because on windows the spwan and process parrallel is using too much memory

  • Release on numba kernel
  • avoid racing when compilation on numba
  • and some other trick
  • fix mp_context and Lock

this replace:
#4331
#4309

@alejoe91 alejoe91 added the sorters Related to sorters module label Jan 20, 2026
node0 = nodes[0]
if isinstance(node0, PeakSource) and node0.need_first_call_before_pipeline:
# See need_first_call_before_pipeline : this trigger numba compilation before the run
node0._first_call_before_pipeline()
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Shouldn't this be a common function (by default pass) for all nodes? I assume any nodes using a numba kernel would benefit from this

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Other nodes are more complicated because we need to run the entire chain.
This help now for peak_detection.
Lets see how we can optimize this

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

sorters Related to sorters module

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants