On Thu, 5 Sep 2002, Christopher Richmond wrote:

>      I wrote a script which continually read in
>       and wrote out a file to disk.  The system
>       ran for about 4 hours before the battery
>       died.    I suppose using the disk so much
>       may have been what caused the system
>       to quit after 4 hours.  Does anyone have
>       any comments about this?  I'm wondering
>       if my battery may need to be replaced.

  Any access to the microdrive, and especially writing, will cause a
serious power drain, so your battery is probably fine.  Hard drives are
the most power hungery, and wireless cards are fairly bad too.  CF cards
are *much* better, but constant access would still lower battery life.

  Unfortunately, NetBSD swap and page behavior can be very bad when trying
to save power.

>       If I suspend the system while running
>       NetBSD with a full battery and leave the
>       system alone for around 12 hours, the battery
>       is down to around 31%.  If I suspend the
>       system while running Windows CE and
>       leave it for the same amount of time, the
>       system is only down to around 97%.

  NetBSD only turns off the display; CE goes into hibernate mode, which
only leaves the memory refresh going.  The problem is that return from
hibernate jumps to ROM code, so in addition to implementing the restart
code you need to disassemble the ROM for each machine to see if you can
trick it into giving you controll back.  This should be possible on all
WinCE machines, but may be somewhat difficult in some cases.  Some Linux
folks have actually done it for at least one machine and found it not too
hard in that case.  I'm going to be working on hpcmips stuff for a while
now and this is on my list, but it isn't at the top.  I don't have a z50,
but if I get it working I will provide detailed instructions that should
allow someone with a z50 to add support without too much difficulty.

Matthew Orgass
darkstar@pgh.net