For comparison of HammerDB TPC-H workloads between systems it is recommended to use the geometric mean of the query times. This can be a straightforward comparison between power tests. When comparing throughput tests it is recommended to compare the geomean of the slowest virtual user when comparing throughput tests of an equal number of virtual users across systems.

0.943
0.332
21.281
1.163
1.93
0.504
13.358
11.419
55.767
20.412
2.435
1.011
2.08
18.064
5.83
2.003
13.503
5.548
1.525
9.765
3.69
3.226
SUM 195.789
GEOMEAN 4.011822724