Current parameter file requests half the available nodes, so we'd like to capture:
- how many nodes were used
- how long the total job took
- what time of day the job began
- total MC simulations run
- volume of ED count data used
- The above information should be written to a separate metadata dir for all jurisdictions, so we can compile a summary report on compute power and time expectations. Parallel or future R packages are likely best for this.
- cronR for macOS/Linux will also produce a script-name.log file of the script run in that dir. We should move these .log files with date they were created to the metadata dir as well
- taskscheduleR for Windows doesn't create log file so it may be best to always capture stderr and stdout for even jobs not scheduled
Current parameter file requests half the available nodes, so we'd like to capture: