Identical systems with widely varying test times

[Isaac]

We have several identically configured systems with large amounts of RAM (512GB) that we are testing. The test times are varying widely with some systems completing the same number of passes in approximately 48 hours and others around 60 hours. The systems are being run with the same settings.

What could account for this?

[David (PassMark)]

Logically something must be different as the process isn't random.

It is just a matter of checking all the possibilities one by one.
Same software version? Same amount of detected RAM? Same RAM speeds? Same CPU clock speeds? Thermal throttling? Same firmware? Same test selection? Zero errors and warnings generated in each case? Same RAM configuration in the same combination of slots? Same MemTest86 configuration file? Same USB drive if you are running from USB? etc...........