- Less aggressive trials of parallel compute - should really scale to size
- Check for JSON switch
Important bug fix release!!
- Fixed bug that had creeped in and always returned the same hash for all runs (terrible)
- Zoomed in on tests to make sure this won't happen again
- Found another potential disaster with too long a command line for parallel jobs
- Move parallel joblog out of the way after a run
- Fixed problem of incomplete cache files by introducing
parallel --joblog - Adding locking so the exact same command does not run twice - with a risk of overwriting output files
- Fetch chromosome names from the annotation file - working of --loco switch changed!
- Use --chromosomes switch to override which chromosomes to compute
- Also allow for parallel GWA compute of non-LOCO (unLOCO?)
- Added metrics page
- Use isolated tmpdir for GEMMA and don't overwrite output files if they appeared during run (make it more transactional)
- Run GEMMA in parallel for LOCO with 5x speedups