Benchmarks
Latency by scheme
and backend.
Compare measured performance across schemes, operations, and hardware backends. Inspect each configuration’s parameters, precision, and execution settings.
CKKS bootstrapping · S2C-first
Latency by backend
Reproduce it ↗Lower is better ↓
AMD Ryzen 9 9950X · Median bootstrap latency
Preset n16_d35_k720_p19_s2cParameters
- Ring dimension N
- 65,536 (2¹⁶)
- Complex values per bootstrap
- 32,768 (2¹⁵)
- Secret Hamming weights
- Dense 1,024 · Sparse 32
- Max. ciphertext modulus
- 1,382 bits
- Max. modulus including keys
- Dense 1,680 · Sparse 120 bits
09.85 s19.7 s29.6 s39.4 s
Select a backend row to inspect its configuration and precision. Speedup is relative to the single-thread reference CPU for the same pipeline.
BackendTransformMedianSpreadSpeedupVersion
Reading the results
Compare the configuration, too.
Compare the same pipeline and preset, and account for the hardware and thread count. Each configuration has 20 measured iterations after three warm-ups. Timings include bootstrapping and return to application scale 2³⁵; setup, key generation, encryption and decryption are excluded.
Latency is the median; spread is median absolute deviation divided by median. Speedup compares each configuration with the single-thread reference CPU for the same pipeline. These results were measured with Poulpy 0.8.3.
The GPU backend is in development, and Binary FHE results are not yet available. Pending results are excluded from rankings and configuration counts.