 Ian Goldberg
|
a9e39d265e
Add a template parameter to RDPF, RDPFPair, RDPFTriple for the leaf width
|
il y a 1 an |
|
Ian Goldberg
|
585e8d16ee
The server shouldn't allocate large amounts of memory when creating RDPFs
|
il y a 1 an |
|
Ian Goldberg
|
3565cedf03
Don't separately account for halftriples that are created as a byproduct of RDPF creation
|
il y a 1 an |
|
Ian Goldberg
|
0fa58457c5
Multithread RDPF creation when it makes sense
|
il y a 1 an |
|
Ian Goldberg
|
94fcf8f8b0
Parallel RDPF evaluator
|
il y a 1 an |
|
Ian Goldberg
|
1216a94599
Avoid default captures in lambda expressions
|
il y a 1 an |
|
Ian Goldberg
|
dfb63bdb5d
Add yield()s in the right places
|
il y a 1 an |
|
Ian Goldberg
|
777f3da468
Remove the unused (and in fact now incorrect) RDPF::size function
|
il y a 1 an |
|
Ian Goldberg
|
704ade665a
Simplify the run_coroutines API when there are a small constant number of coroutines to run
|
il y a 1 an |
|
Ian Goldberg
|
3c86beb097
Change op_counter to aes_ops everywhere
|
il y a 1 an |
|
Ian Goldberg
|
3811dce023
CDPF creation
|
il y a 1 an |
|
Ian Goldberg
|
2e7a8042bd
Refactor RDPF a bit
|
il y a 1 an |
|
Ian Goldberg
|
0de416dfa9
Complete the update protocol
|
il y a 1 an |
|
Ian Goldberg
|
7d614a5eb4
It's faster to do evaluations of RDPFPairs and RDPFTriples in lockstep
|
il y a 1 an |
|
Ian Goldberg
|
473692210f
Add a comment as to why we still have both expand() and Eval::next()
|
il y a 1 an |
|
Ian Goldberg
|
7df8809b6d
Add a DPF streaming evaluator
|
il y a 1 an |
|
Ian Goldberg
|
bc1d850132
Add the ability to pre-expand RDPFs
|
il y a 1 an |
|
Ian Goldberg
|
cac0ca3269
Comment that we use a different technique from Duoram's Appendix D to normalize the unit vector
|
il y a 1 an |
|
Ian Goldberg
|
6d92b7149b
Single-input evaluation of DPFs
|
il y a 1 an |
|
Ian Goldberg
|
b120fc84eb
All parties can now load RDPFs of different depths and keep track of stats
|
il y a 1 an |
|
Ian Goldberg
|
8f0ea5629f
RDPF pairs (for the server) and triples (for the computational peers)
|
il y a 1 an |
|
Ian Goldberg
|
b163fc4270
Templatize the I/O operations for RDPF
|
il y a 1 an |
|
Ian Goldberg
|
c8a0051158
Be able to send RDPFs to other parties
|
il y a 1 an |
|
Ian Goldberg
|
a3baa02860
Be a little more defensive when reading single bits
|
il y a 1 an |
|
Ian Goldberg
|
54556fce46
It turns out you can just use bitwise operators ^ & on __m128i
|
il y a 1 an |
|
Ian Goldberg
|
e9bbb4b09f
Finish generating the RAM DPFs
|
il y a 1 an |
|
Ian Goldberg
|
83d2f82977
Expand on a comment
|
il y a 1 an |
|
Ian Goldberg
|
cfb460526a
Constructing DPFs: the flag correction bits
|
il y a 1 an |
|
Ian Goldberg
|
41ca42d01b
Start constructing DPFs: the correction words
|
il y a 1 an |
|
Ian Goldberg
|
0ac3de8dd2
Count AES operations
|
il y a 1 an |
