We have two processes going, one polls sysprocesses and the other is sysprofiler. Both have measurements for CPU usage in milliseconds. According to documentation CPU usage on sysprocesses is a cummulative total. However, we are seeing CPU milliseconds in the sysprocesses for a SPID that are MUCH (5x) higher than profiler is reporting for the same process. Does anyone have any information that would shed light on why that might happen.