FHE16 Performance Validation

Operation-level benchmark validation for FHE16

Explore FHE16 operation latency and input-range performance through public benchmark charts, category controls, and summary metrics.

View Papers

Interactive Benchmark

Compare latency by operation category

Select a version, execution strategy, runtime target, and operation category. Current numeric data is published only for FHE16 version 0.10 / Latency-first / CPU.

Experiment Computer

Current public results were measured on this computer

When additional machine runs are added, the same selector can switch the chart dataset.

CPU
2 x Intel Xeon Gold 6240R @ 2.40GHz
Cores / Threads
48 cores / 96 threads
Memory
251 GiB RAM
GPU
NVIDIA RTX A5000
ISA
AVX2 / AVX-512
OS
Ubuntu 24.04 · Linux 6.17

Dataset: data/*.json · Current experiment computer run

Execution Profile

Split the benchmark by strategy and runtime target

Latency-first assigns the full thread pool to one operation and focuses on minimizing circuit depth and critical path latency.

Version
Strategy
Target
Version
FHE16 version 0.10
Optimization Focus
Circuit depth and critical path
Parallel Schedule
All threads cooperate on one operation
Runtime Target
CPU baseline using AVX2 / AVX-512 capable workstation execution
Data Status
Published CPU latency dataset

FHE16 benchmark chart

Loading benchmark data...