Why dynamic FPS will eat somany performance
E0x72-21 opened this issue ยท 2 comments
I'm not sure why Spark is attributing this to Dynamic FPS, no mod method calls are even found in the call chain ...
It's normal vanilla behavior for the server to run a tick and then have the thread sleep a bit less than a 20th of a second until the next tick to perform more work. All the LockSupport.parkNanos() method does is park (make the thread sleep / wait) the amount of nanoseconds the called asks for.
Seems Spark also has a page explaining this, note the Generally: A higher sleep percentage is good thing near the bottom.