-
Notifications
You must be signed in to change notification settings - Fork 1
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
#367: make output JSON files compatible with vt offline LB #374
#367: make output JSON files compatible with vt offline LB #374
Conversation
user-defined-memory-toy-problem.mp4
|
This PR is now ready for review; it has actually turned into somewhat of a refactoring, due to the fact that the initial design only considered in-place rebalancing, for the sake of execution speed and memory efficiency. In order to achieve the desired goals of this issue, the underlying model for the |
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.
Will anything in this PR cause a failure for cases where objects can be created or deleted on any phase?
This looks like great progress, but it still does not seem to read in everything in the json that pertains to task attributes or metadata and retain it for output.
Hello dear reviewers @lifflander @nlslatt @pierrepebay this is a reminder :) |
In |
@pierrepebay creating a follow-on issue to fix or deprecate the We can't just remove the one call to the function right now, because it itself calls other functions of that same script. Ideally we would move all of that to a helper class somewhere. |
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.
Ok; other than that I have tested thoroughly the different configurations and gone over the changed features, everything looks good to me.
* #367: first bit of refactoring * #367: internal consistency between writers * #367: improved internal consistency between the 2 types of writers * #367: cleaner use of data parallelism * #367: refactored usage of JSON writer from LBAF * #367: fixed annoying logical error that happened with empty task lists * #367: slight reorg of initializations * #367: whitespace cleanup * #367: further refactoring to prepare for multiphase reading * #367: modified execution flow to allow for multiphase loading * #367: simplified Phase API by removing unnecessary debug info * #367: improved internals and erroring out of reader * #367: further compactification of writer * #367: more renaming with consistent variable names * #367: ensure phase length consistency * #367: removed debug statements * #367: updated reader communications parsing for multi-phase support * #367: completed JSON reader refactoring * #367: whitespace cleanup * #367: attempting different acceptance conf parameters * #367: updated test data with respect to new requirement for empty ranks * #367: renamed stats into data for consistency * #367: overhaul of test harness and data to adapt to code changes * #367: whitespace cleanup * #367: protected against empty task lists in a JSON phase * #367: all tests should now pass * #367: updated Phases test * #367: raise proper exception for failed work model * #367: fixed missing tasks key which for now is required even for empty lists * #367: code style: consistency * #367: better handling of statistics in LBAF * #367: added JSON writer parameters discussed with JL and NS * #367: whitespace cleanup * #367: now passing parameters to writer and optionally compressing * #367: modified LBAF for multi-phase writing * #367: progress checkpoint of refactoring of execution model for multiphase processing * #167: factored out initialization steps * #367: multiphase JSON writer now functional (multi-threaded) * #367: restore Visualizer functionality with new data model * #367: updated PhaseStepper algorithm to new data model * #367: processed --> rebalanced for clarity * #367: whitespace cleanup * #367: fixed typo * #367: upated CI for data model changes * #367: removed debug statements * #367: completed refactoring for multiphase * #367: fixed CI * #367: improved test * #367: better failing tests parameters * #367: fixed rank deep copy bug of shared blocks and testing it now * #367: removed debug statements * #367: modified indexing of processed phase; also improved API * #367: fixed typo * #367: completed implementation with application of load changes * #367: removed debug statement * #367: fixed post-merge errors * #367: fixed an other post-merge error * #367: missing trimmed down data set * #367: fix private variable name * #367: addressed review comment --------- Co-authored-by: Pierre Pébay <[email protected]>
Resolves #367