The PALISADE-CPU backend is a Reference backend engine for HEBench which is a shared library that implements the required functions specified in either the hebench::APIBridge or hebench::cpp wrapper. It is intended only as a reference on how to implement a PALISADE backend for HEBench and it is not fully optimized for performance.
Apache License 2.0
6
stars
2
forks
source link
Current Testing Performance for Eval(Sum|InnerProd) BFV Workloads is lower than Expected #25
The performance for the BFV benchmarks on development utilizing Eval(Sum|InnerProduct) functions are slower than what's currently on main (v0.5.3-beta):
development:
15. matrix_multiplication_10x9_x_9x8_0/wp10 (valbfv) | 162.03 ms | 10.9 s
22. matrix_multiplication_10x9_x_9x8_0/wp10 (eipbfv) | 170.81 ms | 11.35 s
main/v0.5.3-beta:
15. matrix_multiplication_10x9_x_9x8_0/wp10 | 94.73 ms | 6.9 s
22. matrix_multiplication_10x9_x_9x8_0/wp10 | 113.69 ms | 7.5 s
To Reproduce
Steps to reproduce the behavior:
Run with the default input sizes with the commit listed in the versions above, then run with v0.5.3-beta.
Expected behavior
While a bug was fixed in v0.5.3-beta that caused it to fail on some inputs, the expectation isn't that the performance difference should be this great.
Describe the bug
To Reproduce
Steps to reproduce the behavior:
Expected behavior
While a bug was fixed in v0.5.3-beta that caused it to fail on some inputs, the expectation isn't that the performance difference should be this great.
Screenshots
N/A
Additional context
N/A