use proposer index cache for block proposal #4338
Conversation
|
@michaelsproul PTAL, I don't know if I handle some details properly 😂 |
| let (proposers, decision_root, _, fork) = | ||
| compute_proposer_duties_from_head(proposer_epoch, self) | ||
| .map_err(BlockProductionError::BeaconChain)?; |
There was a problem hiding this comment.
I think we could simplify this a little by calling state.get_beacon_proposer_indices on the state directly. It has already been advanced through to the slot of the proposal (above) so it should be able to serve the call.
The compute_proposer_duties_from_head basically does the same thing except it also clones the head again (which is slow). Using the head is also not necessarily correct if we are proposing on a non-head block (sometimes in tests we propose blocks using an old parent instead of the head).
The fork can be set to state.fork() directly, and we also have the dependent_root (same as the decision root) above.
There was a problem hiding this comment.
@michaelsproul updated! But delete some proposer cache prime check in some tests to pass CI.
michaelsproul
added
waiting-on-author
jimmygchen
added
ready-for-review
|
Closing this for now as it doesn't seem to be a promising avenue for optimisation. We can revisit later if necessary. |
Issue Addressed
#4313
Proposed Changes
use proposer index cache for block proposal
Additional Info
NA