Implementing #10102 in #11766 resulted in a performance regression.
This happens because computing the required Chunk::num_events_cumulative_per_unique_time_unsorted takes a lot of time.
We should be able to cache this value and avoid recomputing it each frame, this would make it easier to even more fine grained sampling from chunks and create a better data density graph.
Implementing #10102 in #11766 resulted in a performance regression.
This happens because computing the required
Chunk::num_events_cumulative_per_unique_time_unsortedtakes a lot of time.We should be able to cache this value and avoid recomputing it each frame, this would make it easier to even more fine grained sampling from chunks and create a better data density graph.