Closed aansaarii closed 1 year ago
The PR is now tested on ARM machine, but I didn't observe very obvious difference. @UlisesLuzius should test it in the real environment with 360K to see if it is the problem of negative exponential. Once he tests it and it works, I will merge this PR.
This PR makes two changes: It removes an unnecessary echo line in the client entry point It changes the load generator to use fixed distribution. This way, the actual load is closer to the given RPS.