-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix DIPOL bugs, add quad tests, improve parallelization, other improvements #39
Conversation
improve parallell reduction by setting synchronous = NORMAL journal_mode = WAL
PhotoPolResult.create: pass clear=True to photopolresult.reducedfits.set
include handling of BLLac special case
If I can suggest something for the future, to easier track changes, it is better to solve one bug at a time in single PRs |
I know, sorry. The pace at which we have made changes lately to have a functional pipeline for the latest campaigns has forced us to work at the same time on many issues. |
previous behavior was equivalent to chmod 'a=r', now 'a-w'.
the func modifies the value so subsuquent iterations had different start args it was making some files fail in bulk reduction but not when processed alone in general, it is not recommended to have mutable types as default args
now trying to use missing Instrument definitions will give error
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Overall looks good to me. I just left some minor comments. I think this PR should be merged soon
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
it has not been used in along time and it is not mantained it is not really needed after the last improvements in parallelization
Solves #22 , #38, and some bugs introduced in #23.
Summary of changes:
max_concurrent_threads > 1
, it reduces all polarimetry groups for each epoch in parallel. This reduces time a lot especially in nights with DIPOL data.Concurrency has been improved greatly. Previously there was some isolated 'database errors' with ~10 cores, and many more with 12-14. Now I have tested with up to 16 cores and there are no errors, probably even more cores are fine too.
This is a nice point to create a release (#37).