AnalyticDB:Test Instructions Based on Standard Test Datasets

Test Procedure Overview

The performance test includes the following three core steps:

1. Build data structures: Create table schemas in AnalyticDB for MySQL based on the TPC-H and TPC-DS standards.

2. Initialize data: Generate and import 100 GB of standard test data.

3. Execute standard queries: Run all standard query statements defined by TPC-H and TPC-DS and record the running time.

Execution Policies

To ensure the accuracy of the results, this test follows these execution policies:

• Prefetching: Before the official test, run two full query rounds for system prefetching to warm up components such as the cache and connection pool.

• Multi-round optimization: Run five complete test rounds consecutively and use the best result from the five rounds as the final performance metric.

Test Script Examples

Shell scripts are used to automate query execution and ensure test consistency.

while [ $n -lt 23 ]
do
 echo "query ${n} starting"
 time mysql -f tpch <tpch_${n}.sql
 echo "query $n ended!"
 n=`expr ${n} + 1`
done

while [ $n -lt 100 ]
do
 echo "query ${n} starting"
 time mysql -f tpcds <tpcds_${n}.sql
 echo "query $n ended!"
 n=`expr ${n} + 1`
done

TPC-H Test Results

With 100 GB of data, AnalyticDB for MySQL efficiently executed all 22 complex TPC-H queries (Q1 to Q22). All queries were completed in sub-seconds or seconds, with a total running time of only 11.65 seconds. The following table shows the detailed results in seconds:

TPC-DS Test Results

For the TPC-DS test dataset, which features higher query complexity and more diverse scenarios, AnalyticDB for MySQL also delivered powerful performance. It completed all 99 queries with a total running time of approximately 48.22 seconds. The following table shows the detailed results in seconds: