CPU Affinity

[sonicfromnewyoke]

Problem

Processor affinity takes advantage of the fact that remnants of a process that was run on a given processor may remain in that processor's state (for example, data in the cache memory) after another process was run on that processor. Scheduling a CPU-intensive process that has few interrupts to execute on the same processor may improve its performance by reducing degrading events such as cache misses, but may slow down ordinary programs because they would need to wait for that CPU to become available again

Proposal

Wanted to ask about the vision of further architecture and the possibility of CPU affinity usage.

Are you planning to use the CPU pinning of all Sig's components like it was done by FD team or at least heavily loaded PoH process as it was done in fndn realization ?

...or, maybe, it is already done in the codebase and I just missed it, so appreciate it if you point me out where to look

Alternatives

leave it as it is. But it looks like a potential miss of opportunity to make a more performant solution

Additional Context

P.S. apologizes for the long time wasn't active, all that madness with memecoins... never mind...

[dnut]

We would like to explore this, but it hasn't been prioritized yet. It's not well reflected in our backlog. Currently there is just this one ticket that mentions cpu pinning in accountsdb: https://github.com/orgs/Syndica/projects/2/views/10?pane=issue&itemId=58729449

Some other ideas about multithreading have come up as well. For example we may consolidate some pipelines into fewer threads to reduce the need for thread synchronization.