Kris Bahnsen

@CaptainNic If you got some ideas on how to have an equivalent high accuracy sleep similar to `clock_nanosleep()` in Windows, I'm all ears.

If its not worth the trouble for Win then Its probably fine to sacrifice some accuracy there. Win support is an afterthought for me since the only thing I need...

Dumping some notes here: System idling, built for release with gprof. ``` $ ps -o pid,comm,etime,time -p 22553 PID COMMAND ELAPSED TIME 22553 msemu 04:29 00:01:00 ``` ~0.223 load This...

changes in 831b512c627f18ab4adb696dfce8103587b4aaea ``` ps -o pid,comm,etime,time -p 23178 PID COMMAND ELAPSED TIME 23178 msemu 05:37 00:01:13 ``` ~0.216 Not too much improvement ``` Flat profile: Each sample counts as...

From commit 84bac68c15376e9bbe4e3be048c3a6fda2968e5e ``` $ ps -o pid,comm,etime,time -p 24000 PID COMMAND ELAPSED TIME 24000 msemu 04:24 00:00:41 ``` Roughly ~0.155, or ~28% reduction of CPU time by removing debug...

Rebased against master and this somehow got, better? I also am on a new laptop so there is probably some efficiency gain there. I'll have to reproduce all of the...

Someone mentioned taking a look at zeromq for this task rather than a standard FIFO/named pipe