-
Notifications
You must be signed in to change notification settings - Fork 119
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
building computeKey cache on first fcu
call
#2748
Comments
fcu
callfcu
call
Changing the default here increases the total time it takes to import + compute cache - the way forward here is probably to check that a cache is available on normal (non-import) startup or something similar to that effect, so that it is computed before turning on networking. |
Yes there is slow down, but personally as far as I ran imports for holesky and sepolia. The slowdown is not very significant. But yes the approach for checking on start does makes sense. |
you can measure it to quantify this point, but that only makes sense to do so at mainnet scale, ie around block 20m where the database no longer fits in memory. |
#2788 addresses this somewhat
a significant part of the slowdown is due to the separation of vertex and key data in the database - when the database is big, this leads to a lot of read/write amplification as each mpt node has to be read from two tables - it's worth remembering for future reference however - if we were to join the two tables, it might very well be the case that computing during import is decently fast. |
nimbus-eth1/nimbus/config.nim
Lines 531 to 535 in 0b93236
The default mode should be with validation, will also smooth out the UX from
import
tosyncer
. As with validation the triggering of the computeKey cache won't take place with the firstfcu
call which takes hours and results in thefcu
call being timed outSuggestion:
defaultValue: false
The text was updated successfully, but these errors were encountered: