-
Notifications
You must be signed in to change notification settings - Fork 69
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Remove cpu-effort-percent
and replace with new option
#1784
Comments
cpu-effort-percentage
and replace with new option
cpu-effort-percentage
and replace with new optioncpu-effort-percent
and replace with new option
cpu-effort-percent
and replace with new optioncpu-effort-percent
and replace with new option
I suggest to change the new option from a percentage to an offset in ms for better fine-tuning. The name could be something like “produce-block-offset-ms”. [A general suggestion for new parameter names is starting with the name of the plugin that it relates for easier grouping of related parameters in config.ini and in manual files. E.g. in this case the parameter is for the producer-plugin, so the parameter name starts with "producer-xyz".] |
After discussion on the Node Operators call:
|
…ge meaning to be over complete round
[5.0] Replaced `cpu-effort-percent` with `produce-block-offset-ms`
[5.0 -> main] Replaced `cpu-effort-percent` with `produce-block-offset-ms`
Brian's Edits
Because the behavior of
cpu-effort-percent
has changed significantly, we should create a new option instead.Suggestion noted on changing the default from 90% to 95%, will consider for a future release as we continue to receive feedback from production.
Decisions:
cpu-effort-percent
? No as this is a major release and the impact is limited to failure to start.produce-block-offset-ms
450
msRemaining Tasks
Original Report
@eosusa's original title: "Rename cpu-effort-percentage to round-cpu-effort-percentage and default to 95%"
With the 5.0 RC2 release, the cpu-effort-percentage parameter functionality was dramatically changed from not controlling each individual block's cpu effort during signing, but now controls the ENTIRE round's cpu effort for ALL blocks cumulatively for the round.
While this is a great new feature to add, having it reuse the original parameter from previous versions that functioned differently I believe is a mistake and could easily clear up some confusion by simply renaming the new parameter vs.s reusing the original.
Also, on most public networks we operate on, the default of 90% round-cpu-effort-percentage is probably a bit too low as it kicks out the last block before it's scheduled time, which leads to some confusion because blocks are arriving early (and most networks don't have to send last block that aggressively). We've found most networks tend to function properly with the last block coming out about 300ms early, which would be roughly 95% round-cpu-effort-percentage I believe (if I understand it correctly, that is :))
The text was updated successfully, but these errors were encountered: