diff options
author | Kali Kaneko (leap communications) <kali@leap.se> | 2016-04-08 03:16:18 -0400 |
---|---|---|
committer | Kali Kaneko (leap communications) <kali@leap.se> | 2016-04-08 03:16:18 -0400 |
commit | 9abd16fcbc23306e542aaa5cdc2ea8e9847dbfd7 (patch) | |
tree | e3515203616179e9f81392e3ae6017fe91be70b9 /README.rst | |
parent | 4dec62b52e9ee5982fc3587ea6c8eefc21d76417 (diff) |
add ability to trigger a sync and measure reactor responsiveness
this is bad news though: we're blocking the reactor with the current
implementation.
what's the culprit?? stay tuned...
Diffstat (limited to 'README.rst')
-rw-r--r-- | README.rst | 9 |
1 files changed, 8 insertions, 1 deletions
@@ -58,10 +58,17 @@ c) **Simultaneous easy+load**: Observe how the no-cpu perf case degrades when ru * experiment with different parameters for the process pool. +Run debug soledad server +------------------------- +need patched branch:: + twistd -n web --port 2323 --wsgi leap.soledad.server.debug_local_application_do_not_use + + To-Do -------------- * [x] make the cpu load variable (parameter to fib function: pass it as env var). -* [ ] graph req/sec in response to variable cpu loads (parameter to fib). +* [x] graph req/sec in response to variable cpu loads (parameter to fib). * [ ] graph response of perf-easy DURING a run of perf/perf-little. * [ ] compare the rate of responsiveness against variable cpu loads. * [ ] scale these minimalistic examples to realistic payload decryption using gnupg. + |