Prioritize lower-compute-unit-limit transactions #28791
Conversation
|
This doesn't strictly bump txs with requested CU, but those with lower CU limits. I think in most cases this will result in the same thing, however does not if a tx actually needs higher CUs than the default of 200k. |
| fn cmp(&self, other: &Self) -> Ordering { | ||
| self.priority().cmp(&other.priority()) | ||
| match self.priority().cmp(&other.priority()) { | ||
| Ordering::Equal => other.compute_unit_limit().cmp(&self.compute_unit_limit()), |
There was a problem hiding this comment.
Not fully convinced if it's a good idea to prioritize txs requested less CUs. Maybe better to incentivize user to adapt requested CU with economic means. But if we really want to using prioritization as tool to encourage user to requested CU, then maybe only by a boolean flag of requested vs non-requested. (perhaps better move the discussion to issue itself)
There was a problem hiding this comment.
Does reversing the comparison result in better block packing?
github-actions
bot
added
the
stale
|
See comments above, not the direction we want to go. |
Problem
#28751. Promote transactions with lower requested computer units.
Summary of Changes
Update
OrdforImmutableDeserializedPacketto prioritize lower compute-unit-limit packets after priority.Fixes #