[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: parallelism
From: |
Marcus G. Daniels |
Subject: |
Re: parallelism |
Date: |
27 Mar 2000 17:03:53 -0800 |
User-agent: |
Gnus/5.070084 (Pterodactyl Gnus v0.84) Emacs/20.4 |
The logical model of concurrency does create objects for all instances
of low-level concurrent actions, but it doesn't yet use any operating
system primitives to realize the parallelism based on these objects
(e.g. as threads or MPI sends). This goal is the low level,
implicit side parallelism for Swarm. For example, with the Objective C
`mousetrap' compiled with "make EXTRACPPFLAGS=-DSCHEDULE_INSPECTION",
you can see these objects on the "Mousetrap Schedule" window as counts
on the left and "ActionConcurrent_c" on the right.
The goal of high-level parallelism for Swarm requires tools for
quantitatively determining which sets of agents are
logically/effectively independent and which are interdependent.
We also ought to have better tools for managing experiments
(e.g. testing many sets of initial conditions by running whole
simulations on different CPUs).
The low-level and high-level infrastructure will be different kinds of
technology, but ultimately we want to have the concurrency of a model
evident to the modeler, and to use this information to do things like
smart, automated load balancing.
==================================
Swarm-Modelling is for discussion of Simulation and Modelling techniques
esp. using Swarm. For list administration needs (esp. [un]subscribing),
please send a message to <address@hidden> with "help" in the
body of the message.
==================================
- parallelism, glen e. p. ropella, 2000/03/24
- Re: parallelism, Laurence Clark, 2000/03/27
- Re: parallelism, glen e. p. ropella, 2000/03/27
- Re: parallelism,
Marcus G. Daniels <=
- Re: parallelism, Ralf Stephan, 2000/03/28
- Re: parallelism, glen e. p. ropella, 2000/03/28
- Re: parallelism, Marcus G. Daniels, 2000/03/28
- Re: parallelism, Ralf Stephan, 2000/03/29
- Re: parallelism, glen e. p. ropella, 2000/03/29
- Re: parallelism, Marcus G. Daniels, 2000/03/29
- Re: parallelism, Ralf Stephan, 2000/03/30
- Re: parallelism, glen e. p. ropella, 2000/03/30
- Re: parallelism, Ralf Stephan, 2000/03/31
- Re: parallelism, Marcus G. Daniels, 2000/03/31