NYSE TAQ Benchmark — In-Memory Query Engines

Compare in-memory query engines

Results show query (and data load) performance statistics of various query engines. Filters update the summary charts and detailed comparison below.

Benchmark filters

Solution: All
Additional filters Open source: Yes No Sort columns: Time Sym, Time Hardware: CPU GPU
Threads: Metric: Cold Run Hot Run Second Hot Run
Machine: Size: Data date:
Baseline:
Query Filters
Instrument filter:
Complexity filter:
Include tags:
Exclude tags:

Benchmark summary

Solution Geometric mean[1] of per-query time ratios relative to (lower is better).
Query Failures
number of null execution times (☠)
Wins per solution
on how many of the compared queries the solution posts the fastest time
Max Res Mem (GB)
maximum resident set size of the benchmark process (reported by time -v)
Data Size (GB)
in-memory size of the loaded tables (reported by the engine)

Detailed Comparison

Load Times

[1] A failed query is counted in the geometric mean with a penalty ratio: twice the worst time ratio posted for that query by any of the compared solutions.