125 Commits

Author SHA1 Message Date
Marcel Keller
bf7f8f4b65 Expected communication cost in compiler. 2025-12-24 13:47:42 +11:00
Hidde L
bfe3e7e038 Make interface of Dropout backwards compatible 2025-11-06 16:16:21 -05:00
Hidde L
e17e175b16 Merge FlexDropout into Dropout
Update mnist_full examples using Dropout
Clean up code style
2025-10-21 16:26:15 +02:00
Hidde L
da0e4420d1 Update n_samples and output in test script 2025-10-14 22:38:20 +02:00
Hidde L
d0f955c2d3 avg diff, dropout 2025-10-14 21:43:18 +02:00
Hidde L
f8139db805 Add BERT layers and test script 2025-10-14 19:58:44 +02:00
Marcel Keller
9f2a641a94 Protocols with function-dependent preprocessing. 2025-05-30 13:35:02 +10:00
Vincent Ehrmanntraut
337ba94d06 Fix applyshuffle with dead code elimination and more flexible Matrix secure_permute 2024-12-20 08:32:09 +01:00
Vincent Ehrmanntraut
0db8d7a419 Fix emulation of permutations, use single-instruction matrix shuffle, use multithreading for matrix shuffles 2024-12-19 09:46:58 +01:00
Vincent Ehrmanntraut
5050dc4c51 Eliminated (most) global vars in SecureShuffle, general cleanup 2024-12-14 00:11:22 +01:00
Vincent Ehrmanntraut
0faee7dcb0 Implement shuffle for SpdzWise Rep3 2024-12-12 10:47:23 +01:00
Vincent Ehrmanntraut
684c08277d Fix annoying bug when too many permuatations are performed in parallel 2024-12-12 10:06:04 +01:00
Vincent Ehrmanntraut
70fcbc872c Fix printing of floats 2024-12-12 09:52:59 +01:00
Vincent Ehrmanntraut
d3baedd3b3 Fix merge-induced issues 2024-12-12 09:33:52 +01:00
Vincent Ehrmanntraut
60d481dc4e Merge branch 'refs/heads/master' into parallel_permutations
# Conflicts:
#	Processor/Processor.hpp
#	Protocols/FakeProtocol.h
#	Protocols/Rep3Shuffler.h
#	Protocols/Rep3Shuffler.hpp
#	Protocols/SecureShuffle.h
#	Protocols/SecureShuffle.hpp
2024-12-12 09:25:02 +01:00
Vincent Ehrmanntraut
0f4c825366 Working on parallel matrix shuffle 2024-12-12 08:56:46 +01:00
Vincent Ehrmanntraut
3d3bfcf6a6 Fix calculation of maximal waksman depth 2024-12-11 15:42:04 +01:00
Vincent Ehrmanntraut
3126cfb6c5 SecureShuffle that is able to permute and unpermute parallely, but sequential permutation does not match parallel permutation... 2024-12-11 10:11:59 +01:00
Vincent Ehrmanntraut
01723cc079 Still sequential, but at least applyMultiple works without global vars now 2024-12-11 09:29:05 +01:00
Vincent Ehrmanntraut
b99e2de60f Semi-honest replicated sharing now has parallel shuffling 2024-12-09 11:48:18 +01:00
Vincent Ehrmanntraut
b5467b69a9 Shufflers are now aware of parallel shuffles, but still perform them sequentially 2024-12-09 11:04:28 +01:00
Vincent Ehrmanntraut
b91a01adc3 Make applyshuffle instruction mergeable, execution is still sequential 2024-12-06 15:55:01 +01:00
Marcel Keller
78a4b65933 Allow regint registers as argument in exported functions. 2024-12-04 11:47:08 +11:00
Marcel Keller
91321ff8cd Functionality to call high-level code from C++. 2024-11-21 13:14:54 +11:00
Marcel Keller
d4933d1d30 Update URLs. 2024-07-19 12:31:07 +10:00
Marcel Keller
78fe3d8bad Maintenance. 2024-07-09 12:19:52 +10:00
Vincent Ehrmanntraut
70d65d3c98 Add to_row_matrix and to_column_matrix at array 2024-07-02 12:02:22 +02:00
Vincent Ehrmanntraut
aa2441ca08 Optimize allocator, add timeout for building the dependency graph 2024-06-25 09:15:35 +02:00
Vincent Ehrmanntraut
a490c391be Fix multithreaded MATMULSM, direct_mul. Should also now work with Python 3.8 2024-06-24 11:42:18 +02:00
Vincent Ehrmanntraut
4d7b8a09da Make MATMULSM mergeable 2024-06-11 09:05:19 +02:00
Marcel Keller
ef82a68aa9 Respect -O with TensorFlow compilation. 2024-05-31 11:27:50 +10:00
Marcel Keller
b57dafb66b Fix bug with larger ORAMs in binary circuits. 2024-03-21 16:56:27 +11:00
Marcel Keller
cf4426fdb3 Multinode computation. 2023-12-14 12:17:54 +11:00
Marcel Keller
2813c0ef0f Maintenance. 2023-08-14 18:29:46 +10:00
Marcel Keller
13cd9420f9 Parameter for multithreading. 2023-08-08 15:40:13 +10:00
Marcel Keller
17f683c9ab Conversion of PyTorch regression models with multiple targets. 2023-06-02 20:49:11 +10:00
Marcel Keller
887d9dc916 Merge pull request #1014 from tskovlund/master
Add Path Oblivious Heap
2023-05-26 18:56:59 +10:00
Thomas Skovlund Hansen
b68254305f Print expected outputs in POH tests 2023-05-19 09:21:25 +02:00
Marcel Keller
6cc3fccef0 Maintenance. 2023-05-09 14:50:53 +10:00
Thomas Skovlund Hansen
049833b88f Restructure benchmark to enable more operation rounds than capacity 2023-05-08 11:43:50 +02:00
Thomas Skovlund Hansen
46ccaebba0 Update tests and benchmarks with UniquePOH 2023-05-06 13:04:46 +02:00
Thomas Skovlund Hansen
d40569be86 Change default stash size and accept kwargs in path_oblivious_sort 2023-05-04 07:41:06 +02:00
Thomas Skovlund Hansen
ba7a62fb12 Add more capability to path oblivious sort 2023-05-04 07:04:40 +02:00
Thomas Skovlund Hansen
04dcc12707 Add benchmark parameters N_THREADS and N_PARALLEL 2023-05-03 13:26:43 +02:00
Thomas Skovlund Hansen
8658086cb8 More detailed benchmark printing 2023-05-03 11:01:27 +02:00
Thomas Skovlund Hansen
dd889c3047 Fix init printing in benchmark and format files 2023-05-02 21:48:57 +02:00
Thomas Skovlund Hansen
aa3b492be3 Update PQ benchmark with entry size settings 2023-05-02 21:06:37 +02:00
Thomas Skovlund Hansen
2162e0ddc7 Fix TIME_INIT benchmark bug and set POH debug to false in benchmarks 2023-05-02 13:27:19 +02:00
Thomas Skovlund Hansen
974b36ff40 Add more benchmark functionality 2023-05-02 11:31:19 +02:00
Thomas Skovlund Hansen
dcdc863e8d Add more benchmark functionality 2023-05-01 22:09:43 +02:00