Meeting Notes for 25 Feb

Cosmic

• Replacement of blessing files:
  • STEP 1: we decided to complete these files to update the database with tags for the final group of files that have not been replaced. The new bless and old bless fail the comparisons when we look at the number of lines inside the files and check for the difference in the columns.
  • We decided to add a message on the blessing plots page for the files that have not been replaced.
• Blessing status of files: reports from autobless and the 'dry-run' of the blessing routine and what to do next.
  • We all had a look at the reports from running these routines in data2 and decided to complete the next steps.
  • STEP 2: Edit will update the blessing status of those files with a different blessing status now (as compared to when we first run autoBless.pl). Also, Edit will exclude those that have been analyzed by using a benchmark or are benchmarks (users have changed these files since we ran autoBless.pl).
  • STEP 3: Edit will update the database based on the reports from running the BlessProcessDryRun.java code. Only the files that have a different blessed status now will be updated.
  • When these steps are done, we will be completing the process of replacing the blessfiles for new ones with an improved precision. Yay!!!!
• Documentation:
  • Still working with the recreation of the blessing files. Sudha is working with the Cosmic File Upload diagram.
• Bug in ThresholdTimes:
  • need to recreate static files (Bugzilla 649): Edit will start the process ASAP with a hardcoded cpldfrequency.
• Discussed new requests/bugs to work with:
  • requests for blessing plots (Bugzilla 641 and 645): Edit will add the benchmark fail message to the blessing plots page. The message will only be displayed if user chooses to see it.
  • dynamic plots (Flot vs Plot.ly): We had a look at www13 where there are working versions of these graphs. We are all going to keep testing in order to decide what we are going to use for the dynamic plots.
  • 651: Edit will add code to the 'Add Benchmark' routine. Only those files where we can calculate the duration will be presented as benchmark candidates.
  • Edit will be working with Sudha to get Sudha's computer ready for code development.