Skip to content

NYC Test Performance


With MatrixOne v0.2.0, we follow the Tutorial to execute NYC Taxi test with a standalone server, and compare the results with that of other products, which shows the performance of MatrixOne in query processing time.


The following test results are the average of the three tests in seconds.

Standalone node


The following server was used: AMD EPYC™ Rome CPU 2.6GHz/3.3GHz, 16 physical cores total, 32 GiB Memory.

Query ID MatrixOne 0.2.0 ClickHouse v21.11.4.14 Starrocks v1.19.3
(Only 30M unique data)
Q1 8.37 5.99 2.58
Q2 2.67 4.13 2.18
Q3 3.48 4.56 3.43
Q4 5.34 7.09 4.19
SUM 19.86 21.77 12.38



There are some limitations on quiries in MatrixOne v0.2.0:

  • The filter command WHERE is not surpported
  • The table cannot be partitioned currently, so commands as PARTITION series are not surpported

Learn More

This page shows the results of NYC Test with MatrixOne. For information on other benchmarks that are available when trying out MatrixOne, see the following: