Tests were run on a raw ALMA dataset using the provided CASA calibration script. We compared the time difference between utiliazing the SSDs on Hydra v running the job locally on the spinnning disks of RTDC12.

Hydra steps:
Transfer data to Hydra (not included)
Copy data to SSDs for processing
Run the job on the SSDs
Move the data off the SSDs to Hydra user area.

Conclusion
There was no performance advantage to using the SSDs on Hydra. For ALMA datasets the file sizes are so large that moving the data on and off the SSDs negates any perfomrance advantage. Additionally these tests did not include any extra time to copy the data off Hydra for local examination.

Hydra (not SSD): Test 1 = 1h8m Test 2 = 1h 26m
Hydra (SSD) including copy time: Test 1 = 1h10m Test 2 = 1h4m Tst 3 = 1h18m
Hydra (SSD) excluding copy time: Test 1 = 55m Test 2 = 55m Test 3 = 59m
RTDC12 (HDD): Test 1 = 48m Test 2 = 45m Test3 = 43m