Pablo,

> So Yoshiki, you think that "we" can find some VM workaround for the cache
> problem?... (that's interesting...)
> .. Or maybe the SLC700 got the new chips...

  I think we can't find any workaround for the PXA250 chip anyway...
For PXA255, it sounds that we can get more reasonable performance that
matches the chip spec.

  SL-C700's CPU is known as PXA250.  Hmm.

> I'll try working with the VM... (based on your ARM version -that's the only
> working version i got)

  It would be nice if you can actually take a look at the WinCE VM
code and see any delay or unoptimized primitive or something...

-- Yoshiki