search

Visa-Research / volepsi

Efficient Private Set Intersection base on VOLE
MIT License
98 stars 32 forks source link

question about efficiency between dynamic libraries and the example in fronted #35

Closed jhjiang10 closed 1 year ago

jhjiang10 commented 1 year ago

Hi and thanks for this great work! I have a few questions。 When I use dynamic libraries, I find my speed is nearly twice as slow as the example in fronted. Is that why I use -pic? There are some good solutions。 A million data can be done in about 2s, but it takes about 6s to use the dynamic library

ladnir commented 1 year ago

There should be no difference due to static VS shared. I'm not sure what the issue is. Maybe you are not compiling in release mode.

The idea of Pic is that you can compile a static library and then link this into your own shared library. Your shared library will contain the vole psi static library.

jhjiang10 commented 1 year ago

Thanks！I will try this to do so.

jhjiang10 commented 1 year ago

I have tried to do so and the speed is the same. Thank you for your reply。