At 6:53 PM -0700 9/1/06, Phil M wrote:
I hope that after Universal Binary gets implemented, that Mars (or someone else) can sit down and figure out portions of the Framework which can be Multiple-Processor safe.
I agree that this is becoming more important and could be a big selling point for RB itself and our RB built apps.
I seem to remember hearing that "For Each" is a good possible candidate because of the out-of-order processing.
Yes, that's what they've said. But I wonder how that could really be possible without dramatically overhauling the RB language to add some parallel processing constructs. If it were possible to have multiple "for each" cases running at the same time on separate processors there would be some very thorny problems about coherency of local variables used within the for each loop.
Think about the simple case of incrementing a counter variable to indicate how many for each cases have been processed. If both processors try to increment that variable at roughly the same time there will be a conflict. Both will read the old value, both will increment it and then both will write it back into storage. The variable will go up only by one, even though two iterations have been run.
IMHO it's far more likely that there could be some sort of limited MP thread support for compute intensive operations, and far less likely that For Each will be parallelized across multiple processors.
It would be nice if the IDE itself became MP aware so that compiling multiple targets and possibly even multiple compile phases of the same target could be run in parallel.
Regards, Joe Huber _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html>
