IPC causes spurious awakes of the executor about static_executor

[MartinCornelis2]

Is this fixed / captured by the PR that was just merged, or is this a separate issue?

Additionally, if this is a separate issue, what impact does it have on the executor?
I agree that the executor waking up for no reason is not desirable, but it does not induce unexpected behaviour and I don't expect the performance impact to be major. Can you give me an indication of the urgency of the fix?

Also, what fix would you suggest. Is there another check besides is_ready() that can be performed to prevent this behaviour. And if so, what is the performance impact of the check vs. the performance impact of the unnecessary awakes?

[alsora]

This is a separate issue from the PR created by @mauropasse .
However, that PR already contributes to mitigate the overhead, because it makes sure that no work is performed in these "spurious" awakes.

I agree with you that the impact should no be that much, anyhow, I'm creating some tests for exactly measuring it.

Unfortunately I don't have a solution at the moment, I opened this ticket mainly to make you aware of the issue.
I will investigate more at how the SingleThreadExecutor uses the memory strategy to get ready executables, because that approach apparently fixes the problem. I will see if it's possible to integrate it in the StaticExecutor to gain some performance.