 Ian Goldberg
|
02edb5009c
Relicense the project to MIT
|
il y a 1 an |
|
Ian Goldberg
|
0bb2a5291b
Make the preprocessing communication cost accounting thread-safe
|
il y a 1 an |
|
Ian Goldberg
|
7be5add70e
Floating point error sometimes makes the variance -\epsilon
|
il y a 1 an |
|
Ian Goldberg
|
b906a8da1c
Script to stop all three sets of dockers
|
il y a 1 an |
|
Ian Goldberg
|
f40fbc0dd4
Try a little harder to recover from malformed logs
|
il y a 1 an |
|
Ian Goldberg
|
6ac1f89591
More repro scripting
|
il y a 1 an |
|
Ian Goldberg
|
350433393a
Start on the master repro scripts that control all three docker setups
|
il y a 1 an |
|
Ian Goldberg
|
28931f16fe
Don't load the repro directory into the duoram docker image
|
il y a 1 an |
|
Ian Goldberg
|
e8db098e6d
Correctly compute the batch sizes and numbers for preprocessing
|
il y a 1 an |
|
Ian Goldberg
|
8ab13f5a44
Typo in logfile name to parse for test run
|
il y a 1 an |
|
Ian Goldberg
|
eb38329ad0
Alert if the logs being parsed are malformed
|
il y a 1 an |
|
Ian Goldberg
|
02d8961e96
For the artifact, don't actually write the DPFs to disk
|
il y a 1 an |
|
Ian Goldberg
|
bf8ddd3dcb
Fix missing stanza in 3P preproc log parser
|
il y a 1 an |
|
Ian Goldberg
|
3860ee9051
Add some progress indicators to preprocessing
|
il y a 1 an |
|
Ian Goldberg
|
d3204178f6
Remember to exchange the rotation values in the 2-party online write protocol
|
il y a 1 an |
 avadapal
|
c44a38efeb
some more memory frees
|
il y a 1 an |
|
Ian Goldberg
|
d3566a3e57
We only use 1 DPF for 2-party writes, not 3, so correct the number of OTs we need
|
il y a 1 an |
|
Ian Goldberg
|
e120f6c90e
Correct the amount of RAM used by 2P preprocessing
|
il y a 1 an |
|
avadapal
|
c57db2987e
some memory frees
|
il y a 1 an |
|
Ian Goldberg
|
e9cbf5c3de
Correctly account for Duoram online write costs
|
il y a 1 an |
|
Ian Goldberg
|
465c00d6a5
Found the buffer overflow
|
il y a 1 an |
|
Ian Goldberg
|
e8e9a86321
Output the data points for the figures in the same order as they appear
|
il y a 1 an |
|
Ian Goldberg
|
7083f1f4ed
Better handle incomplete log entries in parse_logs
|
il y a 1 an |
|
Ian Goldberg
|
4271548bb9
parse_logs for Duoram
|
il y a 1 an |
|
Ian Goldberg
|
052de90607
Allow for DUORAM_MAXGB to not be set at all
|
il y a 1 an |
|
Ian Goldberg
|
05b6810824
Large experiments in repro
|
il y a 1 an |
|
Ian Goldberg
|
555d773db0
Touchups to repro
|
il y a 1 an |
|
Ian Goldberg
|
c8fefe83a8
Touch up the run-experiment and repro scripts
|
il y a 1 an |
|
Ian Goldberg
|
6839f3491d
Upgrade the docker scripts
|
il y a 1 an |
|
Ian Goldberg
|
717737ff93
output OT timings and sizes
|
il y a 1 an |
