There was already an optimization in place to not send updates for spawned events regardless of if the hashes were present. This change makes it so fewer of those such events' hashes would be sent in the first place.
This has a slightly negative ~10% impact on CPU performance as the cost of checking for these before sending is slightly higher than the cost of checking after, but it causes a significant reduction in bandwidth use. Up to 40% of event hashes are skipped in the perf test.
There was already an optimization in place to not send updates for spawned events regardless of if the hashes were present. This change makes it so fewer of those such events' hashes would be sent in the first place.
This has a slightly negative ~10% impact on CPU performance as the cost of checking for these before sending is slightly higher than the cost of checking after, but it causes a significant reduction in bandwidth use. Up to 40% of event hashes are skipped in the perf test.