隐语 PSI benchmark 白皮书，新鲜出炉

#sudo apt-get install wgetwget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh#安装bash Miniconda3-latest-Linux-x86_64.sh# 一直按回车然后输入yesplease answer 'yes' or 'no':>>> yes# 选择安装路径, 文件名前加点号表示隐藏文件Miniconda3 will now be installed into this location:>>> ~/.miniconda3# 添加配置信息到 ~/.bashrc文件Do you wish the installer to initialize Miniconda3 by running conda init? [yes|no][no] >>> yes#运行配置信息文件或重启电脑source ~/.bashrc#测试是否安装成功conda --version

# 创建干净的python环境conda create -n sf-benchmark python=3.8
# 进入benchmark 环境conda activate sf-benchmark
# 安装secretflowpip install -U secretflow
# 创建一个sf-benchmark目录mkdir sf-benchmarkcd sf-benchmark

>>> import secretflow as sf>>> sf.init(['alice', 'bob', 'carol'], num_cpus=8, log_to_driver=True)>>> dev = sf.PYU('alice')>>> import numpy as np>>> data = dev(np.random.rand)(3, 4)>>> sf.reveal(data)

RAY_DISABLE_REMOTE_CODE=true \ray start --head --node-ip-address="192.168.0.1" --port="9394" --resources='{"alice": 8}' --include-dashboard=False

RAY_DISABLE_REMOTE_CODE=true \ray start --address="192.168.0.1:9394" --resources='{"bob": 8}'

RAY_DISABLE_REMOTE_CODE=true \ray start --address="192.168.0.1:9394" --resources='{"carol": 8}'

>>> import secretflow as sf>>> sf.init(address='192.168.0.1:9394')>>> alice = sf.PYU('alice')>>> bob = sf.PYU('bob')>>> sf.reveal(alice(lambda x : x)(2))>>> sf.reveal(bob(lambda x : x)(2))

# 生成三份一千万数据python3 generate_psi.py 10000000
# 生成三份一亿数据python3 generate_psi.py 100000000

import sysimport timeimport logging
from absl import appimport spuimport secretflow as sf
# init loglogging.basicConfig(stream=sys.stdout, level=logging.INFO)
# SPU settingscluster_def = {'nodes': [# <<< !!! >>> replace <192.168.0.1:12945> to alice node's local ip & free port        {'party': 'alice', 'id': 'local:0', 'address': '192.168.0.1:12945', 'listen_address': '0.0.0.0:12945'},# <<< !!! >>> replace <192.168.0.2:12946> to bob node's local ip & free port        {'party': 'bob', 'id': 'local:1', 'address': '192.168.0.2:12946', 'listen_address': '0.0.0.0:12946'},# <<< !!! >>> if you need 3pc test, please add node here, for example, add carol as rank 2# {'party': 'carol', 'id': 'local:2', 'address': '127.0.0.1:12347'},    ],'runtime_config': {'protocol': spu.spu_pb2.SEMI2K,'field': spu.spu_pb2.FM128,    },}
def main(_):
# sf init# <<< !!! >>> replace <192.168.0.1:9394> to your ray head    sf.init(address='192.168.0.1:9394')    alice = sf.PYU('alice')    bob = sf.PYU('bob')    carol = sf.PYU('carol')
# <<< !!! >>> replace path to real parties local file path.    input_path = {        alice: '/data/psi_1.csv',        bob: '/data/psi_2.csv',# if run with `ECDH_PSI_3PC`, add carol# carol: '/data/psi_3.csv',    }    output_path = {        alice: '/data/psi_output.csv',        bob: '/data/psi_output.csv',# if run with `ECDH_PSI_3PC`, add carol# carol: '/data/psi_output.csv',    }    select_keys = {        alice: ['id'],        bob: ['id'],# if run with `ECDH_PSI_3PC`, add carol# carol: ['id'],    }    spu = sf.SPU(cluster_def)
# prepare data    start = time.time()
    reports = spu.psi_csv(        key=select_keys,        input_path=input_path,        output_path=output_path,        receiver='alice',  # if `broadcast_result=False`, only receiver can get output file.        protocol='KKRT_PSI_2PC',  # psi protocol        precheck_input=False,  # will cost ext time if set True        sort=False,  # will cost ext time if set True        broadcast_result=False,  # will cost ext time if set True    )    print(f"psi reports: {reports}")    logging.info(f"cost time: {time.time() - start}")
    sf.shutdown()
if __name__ == '__main__':    app.run(main)