Hi, I am currently trying to apply the OpenFHE library to my use case: Vectors of arbitrary length, either integer or floating point (depending on the applicability of bootstrapping) Computation function consisting of multiple multiplications and additions of ciphertexts and plaintexts Apply comp…

Hi! You should totally watch some of the videos from Duality on Youtube. They are @dualitytechnologies796, and you’ll find nice descriptions. I don’t see BGV bootstrapping in the code base (someone correct me if I’m wrong). You can use CKKS pretty well, though. You don’t need to bootstrap after ea…

The current version of OpenFHE includes approximate bootstrapping for CKKS and bootstrapping for binfhe schemes (TFHE and FHEW).

You can find the answer to your question in this thread . Also, I would like to note that OpenFHE includes 128-bit implementation of CKKS which offers higher precision. See the example (src/pke/examples/advanced-real-numbers-128.cpp) for more details.

Thanks for the quick responses @adolgert and @Caesar ! I’ll check out your mentioned approaches. So as far as I understood, Integer Bootstrapping is not yet possible. Maybe it helps if I outline my design a little bit more in detail: First my formula looks like this, where c* represents ciphertext…

The Python doesn’t look right, as you know. It wouldn’t bootstrap with no levels remaining, and if there were no levels after the bootstrap (which can happen for certain choices of parameters) then the bootstrap would inform you that it failed. Also, Python doesn’t truncate 0.250002 to 0.25 when it …

@reneroliveira Can you look into this? Specifically why there is a difference between the C++ and Python versions?

@reneroliveira Looks like the precision used for outputting real-number results in Python should be increased.

@ypolyakov Alright, that’s easy to change. @iquah I’m comparing the c++ and python source code, the main difference is in the GetBootstrapDepth‎ function. C++ is using the following signature: static uint32_t GetBootstrapDepth(const std::vector<uint32_t> &levelBudget, SecretKeyDist secretKeyDist) …

Updates: The bug was fixed in the dev branch and will be merged into the main in the next minor release. Now, that’s how Python output will look like: [image] To test this version, change the repository branch using git checkout dev and the git pull to get the latest development updates.

I figured out that the problem here isn’t printing precision. The file is meant to test bootstrapping, but it constructs ciphertexts that have all of their towers. That’s why GetLevels() returns 0, because no multiplications have happened on these ciphertexts. The example should, instead, create the…

Unlimited computations on same ciphertext with fixed precision requirements

FHE Use Cases

reneroliveira October 31, 2023, 9:21pm 14

Hi @adolgert, thanks for the words about the wrappers, and the interest in helping us. You might like this post!

Topic		Replies	Views
CKKS Bootstrapping during multiplication gives weird output FHE Questions	3	151	November 14, 2024
About performing many bootstrapping FHE Questions questions	6	567	March 13, 2023
Questions on CKKS bootstrapping with some computations after bootstrapping Library Questions	11	1772	June 16, 2023
About precision, scale value and basic operation of boot strapping Library Questions questions , openfhe-help	3	69	November 6, 2025
CKKS Bootstrapping Library Questions	6	972	July 28, 2022

Unlimited computations on same ciphertext with fixed precision requirements

Related topics