Skip to content

Conversation

@tkoskela
Copy link
Member

@tkoskela tkoskela commented Jun 17, 2020

Based on discussion today and in #67 we can't make the velocity state variables completely local 😒 . I reverted PR #95 and then cherry-picked the parts that still make sense. I also added the noise back to the velocity state variables, which was removed in #68, but kept the loop over particles.

Sorry this got quite massive, it contains:

  • Fix handling of velocity state variables
  • Add option to initialise particles to the true initial state
  • Add tests for particle state
  • Make add_random_field thread safe

Fixes #105 and most likely the issues Alex was having Fixes #92

tkoskela and others added 7 commits June 17, 2020 14:09
This reverts commit 2a25a6b, reversing
changes made to 0cb26e8.
Co-authored-by: Mosè Giordano <[email protected]>
1. Update true state on every process to get true observations
2. Calculate weights in parallel from true observations
3. MPI.Gather the weights instead of model observations
@giordano giordano changed the title Tk/revert state split Revert state split Jun 18, 2020
@tkoskela tkoskela marked this pull request as ready for review June 19, 2020 15:08
@tkoskela tkoskela requested a review from giordano June 19, 2020 15:09
Co-authored-by: Mosè Giordano <[email protected]>
@tkoskela tkoskela merged commit c04828d into master Jun 19, 2020
@tkoskela tkoskela deleted the tk/revert-state-split branch June 19, 2020 19:56
@giordano giordano linked an issue Jun 24, 2020 that may be closed by this pull request
@tkoskela tkoskela mentioned this pull request Jun 24, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

3 participants