📊 Detailed Query Results

Exasol vs ClickHouse Performance Comparison on TPC-H SF100

Overview

This report presents the complete query-by-query performance results for 2 database systems tested using the TPC-H benchmark at scale factor 100.

Systems Under Test

Exasol 2025.1.0

Environment & Hardware

  • Cloud: AWS eu-west-1
  • Instance: r6id.8xlarge
  • CPU: Intel(R) Xeon(R) Platinum 8375C CPU @ 2.90GHz (32 vCPUs)
  • Memory: 247.7GB RAM

Software

  • Database: exasol 2025.1.0

Clickhouse 25.9.5.21

Environment & Hardware

  • Cloud: AWS eu-west-1
  • Instance: r6id.8xlarge
  • CPU: Intel(R) Xeon(R) Platinum 8375C CPU @ 2.90GHz (32 vCPUs)
  • Memory: 247.7GB RAM

Software

  • Database: clickhouse 25.9.5.21

Performance Summary

Key Findings

  • exasol was the fastest overall with 238.1ms median runtime
  • clickhouse was 10.7× slower
  • Tested 308 total query executions across 22 different query types

Performance Visualizations

Detailed Analysis

Performance by Query Type

The benchmark results reveal distinct performance characteristics across different query categories:

The following table shows the median performance for each query category across all systems:

Query Type clickhouse exasol Winner
Aggregation 390.00 82.20 exasol
Join-Heavy 5,596.10 178.00 exasol
Complex Analytical 4,894.50 324.70 exasol

Query-by-Query Results

The following table shows the median execution time for each query across all systems:

query baseline_system comparison_system baseline_ms comparison_ms ratio speedup faster
Q01 exasol clickhouse 791.00 2,527.80 3.20 0.31 0
Q02 exasol clickhouse 74.80 1,088.30 14.55 0.07 0
Q03 exasol clickhouse 324.70 4,059.90 12.50 0.08 0
Q04 exasol clickhouse 60.70 2,586.40 42.61 0.02 0
Q05 exasol clickhouse 203.10 8,884.80 43.75 0.02 0
Q06 exasol clickhouse 41.30 164.90 3.99 0.25 0
Q07 exasol clickhouse 265.10 4,894.50 18.46 0.05 0
Q08 exasol clickhouse 73.70 7,908.50 107.31 0.01 0
Q09 exasol clickhouse 942.80 12,545.60 13.31 0.08 0
Q10 exasol clickhouse 557.40 3,408.50 6.11 0.16 0
Q11 exasol clickhouse 144.20 727.30 5.04 0.20 0
Q12 exasol clickhouse 82.20 770.20 9.37 0.11 0
Q13 exasol clickhouse 657.90 5,744.70 8.73 0.11 0
Q14 exasol clickhouse 81.00 245.30 3.03 0.33 0
Q15 exasol clickhouse 379.60 346.40 0.91 1.10 1
Q16 exasol clickhouse 457.70 517.50 1.13 0.88 0
Q17 exasol clickhouse 28.00 5,821.70 207.92 0.00 0
Q18 exasol clickhouse 631.40 5,352.20 8.48 0.12 0
Q19 exasol clickhouse 25.60 2,207.70 86.24 0.01 0
Q20 exasol clickhouse 277.60 390.00 1.40 0.71 0
Q21 exasol clickhouse 384.50 46,707.20 121.48 0.01 0
Q22 exasol clickhouse 94.40 593.90 6.29 0.16 0

Detailed Statistics

The complete performance statistics for all queries and systems:

query system warmup runs median_ms mean_ms std_ms min_ms max_ms
Q01 clickhouse 2,542.80 7 2,527.80 2,520.50 34.00 2,460.10 2,556.80
Q01 exasol 803.50 7 791.00 793.20 5.20 789.50 804.50
Q02 clickhouse 1,308.40 7 1,088.30 1,094.20 25.10 1,071.90 1,146.00
Q02 exasol 94.00 7 74.80 74.80 0.60 74.20 75.60
Q03 clickhouse 4,527.50 7 4,059.90 4,070.00 180.40 3,884.60 4,396.70
Q03 exasol 326.50 7 324.70 325.10 4.90 318.30 333.10
Q04 clickhouse 3,266.70 7 2,586.40 2,617.80 84.40 2,542.80 2,770.40
Q04 exasol 63.30 7 60.70 60.70 0.30 60.30 61.00
Q05 clickhouse 9,544.30 7 8,884.80 8,866.60 117.20 8,720.30 9,080.40
Q05 exasol 267.80 7 203.10 204.50 3.40 202.90 212.20
Q06 clickhouse 996.00 7 164.90 166.00 3.20 162.60 172.10
Q06 exasol 41.90 7 41.30 41.30 0.40 40.80 42.00
Q07 clickhouse 6,395.30 7 4,894.50 4,900.80 29.60 4,876.00 4,963.50
Q07 exasol 267.20 7 265.10 265.10 1.00 264.00 266.70
Q08 clickhouse 7,138.80 7 7,908.50 7,904.90 222.20 7,575.60 8,175.80
Q08 exasol 74.90 7 73.70 77.90 11.60 72.60 104.20
Q09 clickhouse 14,236.30 7 12,545.60 12,998.10 1,180.20 11,949.30 15,509.60
Q09 exasol 943.90 7 942.80 943.20 1.70 941.20 946.50
Q10 clickhouse 5,456.70 7 3,408.50 3,431.60 90.90 3,352.50 3,616.60
Q10 exasol 556.00 7 557.40 558.50 4.50 553.50 565.10
Q11 clickhouse 1,168.00 7 727.30 736.60 42.80 684.60 799.40
Q11 exasol 143.50 7 144.20 145.30 4.20 140.50 153.00
Q12 clickhouse 2,507.00 7 770.20 794.60 46.20 759.90 878.60
Q12 exasol 84.70 7 82.20 82.10 0.20 81.90 82.40
Q13 clickhouse 5,648.20 7 5,744.70 5,712.20 265.30 5,378.10 6,124.70
Q13 exasol 659.60 7 657.90 659.10 3.40 655.40 664.00
Q14 clickhouse 234.50 7 245.30 242.80 4.80 236.70 248.00
Q14 exasol 94.40 7 81.00 80.90 0.20 80.60 81.20
Q15 clickhouse 410.20 7 346.40 352.60 11.20 345.70 374.50
Q15 exasol 376.10 7 379.60 406.70 35.20 377.60 448.20
Q16 clickhouse 538.70 7 517.50 513.00 14.20 484.30 525.30
Q16 exasol 463.80 7 457.70 458.50 6.70 451.30 472.40
Q17 clickhouse 7,380.90 7 5,821.70 6,032.40 528.60 5,580.20 6,897.60
Q17 exasol 29.30 7 28.00 27.90 0.20 27.50 28.20
Q18 clickhouse 5,835.40 7 5,352.20 5,365.80 88.20 5,289.30 5,545.50
Q18 exasol 632.40 7 631.40 632.20 6.30 622.40 640.40
Q19 clickhouse 2,302.70 7 2,207.70 2,216.20 52.60 2,156.60 2,310.90
Q19 exasol 25.70 7 25.60 25.60 0.20 25.20 25.70
Q20 clickhouse 421.60 7 390.00 390.40 6.40 379.20 397.90
Q20 exasol 281.60 7 277.60 277.80 1.80 274.80 280.90
Q21 clickhouse 48,541.10 7 46,707.20 46,802.50 770.30 45,770.50 47,984.70
Q21 exasol 384.30 7 384.50 390.00 14.30 383.80 422.30
Q22 clickhouse 596.30 7 593.90 595.80 16.60 575.70 619.40
Q22 exasol 97.20 7 94.40 94.40 0.30 94.10 95.00

System Rankings

Based on median query execution time:

Rank System Median Runtime Average Runtime Range (Min - Max)
1 Exasol 238.1ms 301.1ms 25.2ms - 946.5ms
2 Clickhouse 2546.6ms 5378.4ms 162.6ms - 47984.7ms

Benchmark Methodology

Workload Configuration

TPC-H Benchmark:

  • Scale Factor: 100
  • Data Format: CSV
  • Data Generator: dbgen

Execution Parameters

  • Warmup Runs: 1
  • Measured Runs: 7
  • Execution Mode: Sequential (single connection)
  • Metric Reported: Median execution time

Performance Measurement

All queries were executed with the same data and parameters across all systems. The median execution time from 7 runs is reported for each query to minimize the impact of system variance and outliers.

Conclusion

This benchmark provides a detailed, query-level comparison of 2 database systems on analytical workloads. The results demonstrate the performance characteristics and trade-offs of each system when processing TPC-H queries.

While exasol demonstrated the strongest overall performance in this test, the optimal choice for a specific use case depends on multiple factors including workload characteristics, operational requirements, and system integration needs.

Query Details