In general, these type of things either work or they don't. A quick 10 - 15 minute test would be more than sufficient. Testing for a longer period (say multiple hours) really won't give any better idea of the long term reliability, so I'm all for a quick test of some subset of the batch and then ship them all out.
Waiting two days for testing doesn't make much sense, as that impacts everyone, whereas if I did receive a bad one, then only I would be not hashing for the time it takes for me alone to get a replacement, rather than no one being able to hash while all of them are being tested.