Details
-
Task
-
Status: Closed
-
Major
-
Resolution: Abandoned
-
None
-
None
Description
The scripts for running the SystemML performance suite (see SYSML-38) are currently only semi-automated. A fair amount of manual intervention is required to run the algorithms tests from beginning to end. This task covers automating the performance testing process to the point where a single command line will:
- Generate data
- Run all algorithms multiple times at multiple data sizes
- Generate a full set of reports and graphs
- Log any failures and either continue or stop after each failure