Open Ldisi opened 5 years ago
I've got the same problem as yours?Could I ask you some questions? How do I expand the modulus?(if the size of modulus is limit to the precision of long?)
t
, say K > 1
co-prime t
.K
keys.K
copies of ciphertexts. K
results using the CRT.
I've been using HElib for a while,However, I was stumped by the difficulty of data accuracy. I mean , Helib allows to encrypt positive integer only, but we may need the decimal part in practice. My first thought is to scale the encrypted data with a number , maybe 1000 or bigger . Unfortunately, with the augment of data ,I have to increase p ,which leads to smaller L (Number of multiplication) .The logic is that : decimal -> scale data -> lager p -> smaller L . Do you have any good Suggestions for my problems? Maybe you also have insights in division. Happy for advice from everyone!