Google's fully homomorphic encryption package

Posted Jun 15, 2021 2:33 UTC (Tue) by ras (subscriber, #33059)
In reply to: Google's fully homomorphic encryption package by floppus
Parent article: Google's fully homomorphic encryption package

I don't see how it jibes with 0.1 either, but 13ms can be related to the 13ps switching speed of 14nm CMOS [0]. It's a billion times slower. I guess you won't be using it to process long strings.

[0] Clearly my 13ps number should be taken with a bag of salt. Accuracy wasn't the only consideration when choosing it. More accurate figures can be found in Table II here: https://acadpubl.eu/jsi/2018-118-18/articles/18e/3.pdf