Efficient Batched Oblivious PRF with Applications to Private Set Intersection (CCS 2016)
C++ Makefile
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Batched Oblivious PRF

This is the implementation of our CCS 2016 paper: Efficient Batched Oblivious PRF with Applications to Private Set Intersection[ePrint].

Evaluating on a single server (2 36-cores Intel Xeon CPU E5-2699 v3 @ 2.30GHz and 256GB of RAM) with a single thread per party, ours protocol requires only 3.8 seconds to securely compute the intersection of 2^20-size sets, regardless of the bit length of the items.


Required libraries

C++ compiler with C++11 support. There are several library dependencies including Boost, Crypto++, Miracl, and Mpir. Our code has been tested on both Windows (Microsoft Visual Studio) and Linux. To install the required libraries:

  • windows: open PowerShell, cd ./thirdparty, and .\all_win.ps1
  • linux: cd ./thirdparty, and bash .\all_linux.get.

Some Compiling Issues & How to Fix (raised by users):

  1. Problem with compiling mpir: dowgrade sed to version 4.2.2-8 (for Debian Sid). Read more here
  2. Run this project with version>=6 of g++: add static casts (static_cast<int>(0x...)) to lines 27-34 in wake.cpp of crypto++
  3. Error with _mm_cvtsi64_si128: try on a 64 bit system

Building the Project

After cloning project from git,

  1. build bOPRFlib project
  2. add argument for bOPRFmain project (for example: -t)
  3. run bOPRFmain project
  1. make
  2. for test: ./Release/bOPRFmain.exe -t


Our database is generated randomly. We have 2 functions:

1. Unit Test:

test PSI result for a small number of inputs (2^12), shows whether the program computes a right PSI. This test runs on one terminal:

./bOPRFmain.exe -t

2. Simulation:

Using two terminals, compute PSI in 6 cases with the number of input (2^8, 2^12, 2^16, 2^20, 2^24). For each case, we run the code 10 times to compute PSI. The outputs include the average online/offline/total runtime (displayed on the screen) and the output.txt file. Note that these parameters can be customized in the code. On the Sender's terminal, run:

./bOPRFmain.exe -r 0

On the Receiver's terminal, run:

./bOPRFmain.exe -r 1


Our code utilizes some parts of:

We would like to thank all the users for pointing out the compiling bugs that appear on different operation systems and supporting us to fix them

For computing 2-party PSI with NO stash bins, we refer to efficient libPSI.


For any questions on building or running the library, please contact Ni Trieu at trieun at oregonstate dot edu