we want to reproduce the single epoch logistic regression training from the mp-spdz paper (source code). from the paper, the high-level description of the benchmark: "one epoch of logistic regression for 1280 examples of 10 features and batch size 128".
we will just unroll the loop required to do this, which should hurt our total compilation time relative to MP-SPDZ, but not our total run time
we want to reproduce the single epoch logistic regression training from the mp-spdz paper (source code). from the paper, the high-level description of the benchmark: "one epoch of logistic regression for 1280 examples of 10 features and batch size 128".
we will just unroll the loop required to do this, which should hurt our total compilation time relative to MP-SPDZ, but not our total run time