 Ian Goldberg
|
1ae94b11ed
Rename ecall_authenticate to ecall_ingestion_authenticate
|
1 year ago |
|
Ian Goldberg
|
a4c89b33b2
In send_client_mailbox, send the tokens and mailboxes to each client in a single async_write call
|
1 year ago |
|
Ian Goldberg
|
ce01061057
Allocate space for the priority field in public routing
|
1 year ago |
|
Ian Goldberg
|
c178ad1c60
Touch up Client/clients.cpp
|
1 year ago |
|
Ian Goldberg
|
faa016dfb7
Touch up Client/clientlaunch
|
1 year ago |
|
Ian Goldberg
|
46e6ab74a4
Minor touchups to App/start.cpp
|
1 year ago |
|
Ian Goldberg
|
1b285acd68
What was called epoch_duration is really epoch_wait_time
|
1 year ago |
|
Ian Goldberg
|
9302a275e5
Touch up App/net.{cpp,hpp}
|
1 year ago |
|
Ian Goldberg
|
7e7544439b
Wrap some long lines in App/teems.cpp
|
1 year ago |
|
Ian Goldberg
|
5d5f9d9071
Touch up App/appconfig.{cpp,hpp}
|
1 year ago |
|
Ian Goldberg
|
6f8e0dac0e
Clean up the python in App/launch a bit
|
1 year ago |
|
Ian Goldberg
|
4b2813f794
Don't put the Enclave.config.xml file in git, since run_experiments creates it
|
1 year ago |
|
Ian Goldberg
|
fcd39fd3a5
Add TRACE_SOCKIO tracing for client<->storage server connections
|
1 year ago |
|
Ian Goldberg
|
4139cd49db
Add integer time data points to the core-time tradeoff output
|
1 year ago |
|
Ian Goldberg
|
9ae699f39c
A script to compute the core-time tradeoff graph
|
1 year ago |
|
Ian Goldberg
|
1c6525f520
Clean up the log parser and have it record Waksman network creation times as well
|
1 year ago |
|
Ian Goldberg
|
5313a77081
Automatically choose the horizontal scale for the trace plotter
|
1 year ago |
|
Ian Goldberg
|
84be84deee
Experiments with small numbers of single-core servers (<= 3) for 2^20 clients don't work
|
1 year ago |
|
Ian Goldberg
|
4776d3ee86
Fix off-by-one error in gen_manifest
|
1 year ago |
|
Ian Goldberg
|
ccbb4a4dbe
Add 3 and 6 server configurations to the experiment list
|
1 year ago |
|
Ian Goldberg
|
17a1433f5c
Use only 2 cores for 2 or 3 servers
|
1 year ago |
|
Ian Goldberg
|
4c7ec57bb3
Add extra time for m=3 servers as well as for m=2
|
1 year ago |
|
Ian Goldberg
|
a7720fbad8
When using M=1 nodes, there's no communication, so set min_ts based on other reported fields
|
1 year ago |
|
Ian Goldberg
|
c1069ecd2f
Start a couple of extra async io threads
|
1 year ago |
|
Ian Goldberg
|
1b774a6160
Have run_experiments print whether it's doing public or private routing
|
1 year ago |
|
Ian Goldberg
|
87a8e1344a
Don't forget the numactl on the clients
|
1 year ago |
|
Ian Goldberg
|
2f6d22da17
More sleep timing touchups
|
1 year ago |
|
Ian Goldberg
|
65c2cb68db
Log and plot the times ingestion starts and ends
|
1 year ago |
|
Ian Goldberg
|
d2b97fef39
Plot the epochs, rounds, and precomputation phases
|
1 year ago |
|
Ian Goldberg
|
9471a0b20e
Trace and plot node-to-node messages
|
1 year ago |
