not hoisting loads/stores

[Dan Nicolaescu]

Jeffrey A Law <law@cygnus.com> writes:

  >   In message < 199912041331.aa15618@gremlin-relay.ics.uci.edu >you write:
  >   > Hi!
  >   > 
  >   > Here is a simple example where gcc does not hoist loads/stores out of
  >   > a loop. 
  >   > 
  >   > I don't see any reason why it shouldn't do it... 
  >
  > Because those instructions are created long after the loop
  > optimizer has run and thus the compiler never has a chance to
  > hoist those memory refereces out of the loop.

Is there any chance for the loop optimizer to hoist whatever insns are
there (and that will later become those memory references) at the time
it is run? 
Is this a problem with the loop optimizer? 

dan