Nov 5, 2024

So I theorized that there is some fair scheduling going on based partially on quantity, not just priority (which of course is necessary), causing the confirmer to fall behind the validator despite having less work, because there is only 1 confirmer dispatch thread vs. 64 dispatch/work threads for the validator. Otherwise I would expect the confirmer to never fall behind the maximum validation gap, which should rarely exceed the download window (concurrent downloads wait until the last gap is filled within the download window). So with a 10k block download window I expect to rarely see confirmation more than 10k blocks behind validation.