-
Notifications
You must be signed in to change notification settings - Fork 16
Microbenchmark
Patrick Jattke edited this page Jan 5, 2021
·
2 revisions
The microbenchmark
evaluates the performance of the native operations that the FHE schemes support, i.e., addition, multiplication, encryption (symmetric/asymmetric), decryption, and rotations (if applicable). These operations are tested between ciphertexts (ct, ct) and between ciphertext/plaintext (ct, pt). Where available, we also include in-place variants of these operations, e.g., a = a + b instead of c = a + b, which involves creating a new ciphertext c.
We report the time it takes to execute an operation, averaged over 100 repetitions, for the following tools/schemes:
- SEAL BFV
- SEAL CKKS
- PALISADE BFV/BGV[^1]
- PALISADE CKKS
- PALISADE FHEW
[^1]: BGV only used for rotations as PALISADE's BFV implementation does not support rotations.
- Home
- Compilers & Optimizations
- Libraries
- Benchmark Programs
- Implementation Docs