Background of this task: Currently, small queries (e.g. scan 2 million rows and do aggregation) can not fully use CPU resources on 8 logical core machines (e.g. r5.2xlarge).
Also, consider making the default thread pool size to be 2x of CPU number. And check whether it is more efficient for small queries.
Background of this task: Currently, small queries (e.g. scan 2 million rows and do aggregation) can not fully use CPU resources on 8 logical core machines (e.g. r5.2xlarge).
Also, consider making the default thread pool size to be 2x of CPU number. And check whether it is more efficient for small queries.