Hi, I'm attempting to verify that the performance of a new server that we are planning to roll out is as good as or better than the existing server. I used profiler to capture a trace and replayed it against the new server. For some of the key stored procedures, the average cpu per execution was significantly higher by up to 20x. Also, the number of reads was about 5x greater on the new server. I used SQL Compare to verify that all of the procs, tables, indexes and stats were the same and that the record counts were also the same. When I reviewed the perfmon data I saw that as soon as I start the test the number of page faults/s goes up to around 20,000 and the page life expectancy drops to around 5. When I stop the test the page faults falls to almost 0 and the page life expectancy immediately begins to increase. Can anyone explain what could cause this behaviour or what other performance counters I should look at to try to identify the cause of the problem. Thanks!!! Tim.