Oriel — AI & technology worth noticing, every day.

Remove pgrust: the entry is cheating

Repository · https://github.com/clickhouse/clickbench

First seen by Oriel on .

AI summary

ClickBench benchmarks analytical databases on a production-derived dataset and 43 queries, emphasizing reproducibility while documenting substantial limits on how results can be compared.

  • The workload targets clickstream analysis, web analytics, logs, and event data, using an anonymized dataset of 99,997,497 records in a single flat table.
  • Its documented, semi-automated process uses inexpensive cloud VMs; tests can take about 20 minutes for some systems and several hours for others.
  • Queries run three times: the first run is cold, while the faster successful second or third run is used for hot results.
  • The benchmark runs queries sequentially and does not measure concurrent query capacity in its main sweep; a supplementary concurrent-QPS test is run for supported setups.
  • Results are limited by the dataset model and scale, hardware and system differences, and occasional variability; the authors caution against treating scoreboards as definitive rankings.

#database benchmarks #olap #sql #performance testing

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Points

Score reported by the source.

  • Hacker News: 190, latest observation ; +183 over observed period since

Comments

Comment count reported by the source.

  • Hacker News: 148, latest observation ; +147 over observed period since

Aliases

SourceExternal IDURL
Hacker News50205816https://github.com/ClickHouse/ClickBench/pull/2165 Read discussion