Hi everyone,

Do Windows CE devices (in particular the EG-800 MIPS 4122 PocketPC) ramp the
processor clock speed up and down depending on the current processor load? I
imagine they might do this to conserve battery power.

If that is the case, is it possible to control the ramping and force the
device to the maximum frequency?

I am finding that I get spurious performance results depending on how loaded
the processor is prior to executing a particular task. The application is a
custom C++ application originally using ADOCE to SQLServer2.0 now using
OLEDB. The performance degradation does not seem to be database specific and
generally occurs in all areas of code for a particular task.

An example of this scenario is: -

I need to perform tasks A, B, C, D and E. If I perform them in
succession task C takes 20 seconds. If I just perform task C it takes a
whopping 80 seconds.

Thanks in advance.

Regards,

David.