socket.monitor() low interval = high cpu utilization
Opened this issue · 0 comments
tcf909 commented
Hello,
When running 10's of sockets with a monitoring interval of 10ms the cpu utilization pegs a single core (no other work being done [idle event loop]).
Moving the interval to 1000 brings the cpu utilization down, but still seems higher than it should be for an idle state (60+%).
I noticed you changed the polling to timers in a prior issue to eliminate cpu utilization issue. Wondering, why I might be running in to this again in a multi-socket scenario.
node-zmq: 2.15.3
libzmq: 4.1.5