Skip to content

Project Meeting 2022.12.01

Vivek Yadav edited this page Dec 8, 2022 · 5 revisions

Agenda

Task and Code Review Updates

Memory Profiling and Data Types

  • WSP tested the ARC 25% run’ mandatory tour scheduling model, work tour 1 choosers table.
  • WSP suggested converting all string variables into Enums, creating a data model (issue #617), and using data types with lower memory footprints.
  • RSG suggested reviewing the implementation in the mandatory tour scheduling YAML file. It has the option to simulate choosers columns which lists the columns that are included in the choosers table. Some of these columns can be eliminated which can reduce the size of the data table.
  • Wu suggested doing some minor changes that could reduce the memory footprint, especially for the columns that are object data type.
  • Stephen informed about the performance increase in python 3.11. He tested the model with python 3.11 but couldn’t run the model due to lack of updated dependencies in sharrow.
  • WSP highlighted a potential memory leak issue in the mandatory tour scheduling model.
Clone this wiki locally