> Isn't that the same?
Yes, I think so -- I didn't realize:
> Sfx only executes unconditionally when the rate of the statement is less
> than the rate of the enclosing guard.
When I was constructing the example. From Giorgio's response, though, I
think that by either leveraging:
> You'd have to put i-rate statements in
> opcode to get them to execute uncoditionally.
Or by changing the guard in foo to be a-rate (by making a dummy asig
variable and ||'ing it with the k-rate expr in the guard), its possible
to create the example that exposes the current differences in SAINT and
Sfx.
Yes, I'm bored by beating this dead horse too, but I think its actually
relevent, because if in fact there's no SAOL program that exposes
the difference between "uncoditional" and "conditional" approaches to
these slower statements in a UDO, then it really _doesn't_ need to
be nailed down in a corrigenda one way or the other ... intuitively,
though, it would seem like it would have to be possible to write a SAOL
program that exposes the difference.
--jl
-------------------------------------------------------------------------
John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
lazzaro [at] cs [dot] berkeley [dot] edu www.cs.berkeley.edu/~lazzaro
-------------------------------------------------------------------------
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 12:03:56 EST