piranha
piranha copied to clipboard
ucbrise
Piranha: A GPU Platform for Secure Computation
Hi there! I am currently trying to run piranha, however, it gives me failure (segfault) on unit test case @ FuncTest/2.Reconstruct. Due to this seg fault, nvprof also doesn't work....
colabで実行すると,makeする際,valueの値を用いているためmake出来ないとエラーがでた.以下にエラーの文の一部を出力する. /usr/local/cuda/bin/nvcc -dc -Xcompiler="-O3,-w,-std=c++14,-pthread,-msse4.1,-maes,-msse2,-mpclmul,-fpermissive,-fpic,-pthread" -Xcudafe "--diag_suppress=declared_but_not_referenced" -Xcompiler="-DFLOAT_PRECISION=8 -DTWOPC" -c src/main.cu -o build/main.o -I '/usr/local/cuda-11.5/include' -I 'ext/cutlass/include' -I 'ext/cutlass/tools/util/include' -I 'include' /usr/local/cuda/bin/nvcc -dc -Xcompiler="-O3,-w,-std=c++14,-pthread,-msse4.1,-maes,-msse2,-mpclmul,-fpermissive,-fpic,-pthread" -Xcudafe "--diag_suppress=declared_but_not_referenced" -Xcompiler="-DFLOAT_PRECISION=8 -DTWOPC" -c src/nn/AveragepoolLayer.cu -o...
Hi, I read your platform is protocol-agnostic, but the paper mentions you implement and evaluate only semi-honest protocols. Does your platform also support malicious protocols? Thanks
Hi, I would like to better understand the tradeoffs between using built-in floating point kernels like in CryptGpu and utilizing the custom integers kernels of Piranha. I know there is...
The piranha usenix paper states that the comparison protocol of P-Falcon is [Rabbit (Makri et al)](https://eprint.iacr.org/2021/119). I'm quite interested in it, but I'm not able to find the code that...
We noticed that the shape of data samples in CIFAR 10 is 32x32x3, while for MNIST is 28x28x1. Moreover, the number of parameters in AlexNet is much bigger than LeNet....
Hey. I try to run the experiments for LeNet model. However i noticed that the convergence of LeNet is weird. The test accuracy of LeNet remains low (far from 98%)....
There is a small wrong with the implementation of Mult, And and Matmul in the TPC protocol: the client did not add the result of local computation.
Hello, Insofar as I can tell, in the 2PC code, shares of beaver triples are generated as (0, 0, 0) for both parties. Due to this (and some other code),...
