Database test: dual Intel Xeon 5160

Now that we have looked at the prefetchers as well as the kernel versions, it is time to pit the two machines against each other. To be clear, we installed the new kernel version on both servers, and in both the Dell and the Fujitsu server, the prefetchers have been switched on. Below, we show the effect of using a server with a processor that has been clocked 12.5% higher as well as having 25% of additional memory bandwidth. As usual, we only show results of heavy loads (25+ users).

2.6GHz/533 vs. 3GHz/667	MySQL 4.1.20	MySQL 5.0.20a	PostgreSQL 8.2-dev
1x single-core	+13.3%	+12.8%	+9.2%
2x single-core	+11.3%	+14.9%	+11.1%
1x dual-core	+13.4%	+18.8%	+14.3%
2x dual-core	+10.1%	+10.2%	+9.9%
Average	+12.1%	+14.1%	+11.1%