>>>>> mheavner <miheav...@gmail.com> (m) wrote: >m> I realize that the Queue would be the best way of doing this, however >m> that involves transferring the huge amount of data for each call - my >m> hope was to transfer it once and have it remain in memory for the >m> subprocess across run() calls.
Which huge amount of data? The datastructure you talked about can remain in the process. You only have to transfer the input for your calculation in the queue but you have to do that anyway. And there is only one run call per process. When run has terminated the process exits, so you would have a loop in the run(0 method getting work from the queue. -- Piet van Oostrum <p...@cs.uu.nl> URL: http://pietvanoostrum.com [PGP 8DAE142BE17999C4] Private email: p...@vanoostrum.org -- http://mail.python.org/mailman/listinfo/python-list
