A better solution ...

matthew_heusser at mcgraw-hill.com matthew_heusser at mcgraw-hill.com
Fri Jan 26 12:04:41 CST 2001




For all who attended my presentation:

  For those who complained about network bandwidth and
wasted clock cycles ...

  This wasn't a very good application to demonstrate
MPT-Spaces.  The best kind of applications are ones
where the server can transmit a very small amount of data,
and have the client work on that data for long periods of time.

  10 clients at a time * 1 hour to find an answer = 36,000
simultaneous connections.

    A better solution would be to have the master Assign work
(IE "Do 1000 operations") and have the client programs create
an evaluate 1000 darts, then write back totals ("Results1, 650, 350")
 Where the first variable is the number of TRUE and the second
variable is the number of FALSE.

  The master just needs to grab all the results and total them.

  In any event, it was a good proof-of-concept.  Now on to getting
it competitive with Java ...

regards,

Matt H.






More information about the grand-rapids-pm-list mailing list