Karol Sobczak

# Description Large dynamic filters might be intersected with other large dynamic filters (e.g. when collecting DFs and in subquery cache context). Hence performance of such operations is important: ```...

Make benchmarks hardware agnostic

8

Currently benchmarks contains attributes which are specific to the hardware and cluster: - data size (10GB, 100GB, 1TB) - number of runs, number of prewarms - file format This makes...

Hi, Recently I have been playing around with codegen and its parameters. I have changed deathtime and ttarg values and also decaying speed. I have managed to improve echoprint accuracy...

example: ``` @Inject(name="variable") String variable ``` configuration: ``` variable=1234 ``` In such case `variable` would be null. However if in configuration `variable=1234t`, then `variable` is injected correctly. See: https://github.com/Teradata/presto/pull/338#issuecomment-246400538

Partial aggregation needs to produce symbols requried by join equi-conditions and filter expression

Join reordering stats11 speedup

1

Using dedicated cache for join stats ``` After: Benchmark (joinReorderingStrategy) (numberOfTables) Mode Cnt Score Error Units BenchmarkReorderJoinsConnectedGraph.benchmarkReorderJoins COST_BASED 10 avgt 30 5814,628 ± 188,457 ms/op Before: Benchmark (joinReorderingStrategy) (numberOfTables) Mode...

Subquery cache & friends

7

Implement subquery cache for Hive/Iceberg/Delta Subquery cache is a lightweight mechanism for caching source stage computations. It works across queries, but also within a query if similar subqueries are identified....

https://github.com/trinodb/trino/actions/runs/8986058614/job/24681403065 Might be related to https://github.com/trinodb/trino/commit/c25ab366bdb08d7a0813c7cff9c521e83633ee48 fyi: @ebyhr ``` [INFO] Run 255: PASS [INFO] Run 256: PASS [INFO] Run 257: PASS [INFO] Run 258: PASS [INFO] Run 259: PASS [INFO]...

Improve stats reporting for group by operator

2

Similarly as in https://github.com/trinodb/trino/commit/f46fd9c1dccb975bf480f8c7029daf7e45541b54 we could separately report hash lookups/updates and aggregations' accumulator updates to get more insight into query bottlenecks. cc @dain @raunaqmorarka