I am hoping maybe one of you can point me in the right direction. I have
written an application and a device driver for reloading the flash on a piece
of hardware we use for out-of-band mangement. When I load the application
and driver on a server running W2K 64-bit edition, it works fine. However
the group using my app and driver plan to do so from a 64-bit edition of
WinPE 1.2 image running straight out of RAM. I thought this would be fine,
and the app and driver do work, but they run 1/10 of the speed. Even on the
exact same box.
In other words doing a firmware flash of my device on a box running the full
blown OS takes 2 mintues, and running the exact same code on the exact same
box only booted to PE takes over 10 minutes. Why is that? I was told it is
the same kernel. Maybe this is incorrect? Should I be compiling the driver
against the XP DDK or something? Help! There must be something I can do in
my driver or app to minimize this. I've tried obvious things like limiting
screen paints and loading all my firmware binary to memory before pushing it
to the device.
So far nothing seems to make any difference. I'd appreciate any insight
some of you experienced with the pre-execution environment might be able to
provide me.