RE: objective c?

[Miles Parker]

>>> "Stumpo, Cody (C.M.)" <address@hidden> 12/07/99 01:01PM >>>
Where would people say the java speed as a percentage of obj-c speed is now
with swarm?  Roughly, it seems like about 25% to me.  

I think its prob. a little higher then that. Note that this is very dependent 
on apps, but I've thought approaching 40-50% or so. Of course, thats on an NT 
box w/a fast JVM, adn running swarm under Cygnus, so YMMV. But, as Marcus 
suggests, this is (my intepretation) very largely a function of the glue 
etc..at this point, there is really little reason (except perhaps for limited 
graphics stuff) to think that Java would be slower than ObjC, and mayn reason 
to think it would be faster in certain areas.

My own 'pure Java' models typically run at least at comparable speed, and 
dependent on the model can be significantly faster than equivalent swarm 
models, and there is no reason to expect that that wouldn't be the case with 
Swarm as more of the engine is placed natively in Java.

BTW, I n general I strongly agree with the approach of making the implentation 
itself faster (through more Java pieces, or parralelization, or whatever) 
rather than resorting to compiler hacks and the like. 

While I'm at it, Glen's comments re: implementing in multiple languages (and 
I'd add, environments :-)) is dead on. Finally, I'd again make the strong claim 
that there is very little to no 'truly' dynamic behavior that you lose from 
going to Java that cannot be closely replicated in Java given a free design 
hand, but I'd like to hear arguments to the contrary.




-----Original Message-----
From: address@hidden [mailto:address@hidden 
Sent: Tuesday, December 07, 1999 12:41 AM

PJ> I mean, post in the list about where you think we
PJ> might be in 6 months or 1 year in the future and how the java
PJ> thing will be.

Performance-wise, I think the Java layer (or any other languages layered on
the single-threaded Objective C activity library), might get to be 80%
as fast as Objective C models on the Objective C activity library.  To get
past that it would be necessary to unify Objective C object representations
with gcj Java/C++ object representations (major compiler hacking).  Such
a project would be way more time consuming than just re-implementing the
activity library in Java or C++.

My opinion is that, provided everything Java-wise *works* (and there is
certainly room for improvement), I don't really think it is a big deal
if it is, say, half as fast.  A longstanding goal for Swarm, is to
realize a scalable implementation of the logical concurrency model so
that events having the same time run on different CPUs in parallel.

Unless SDG members really want a singleminded focus on the problem
of making the existing virtual machine (i.e. activity library) faster with
the Java layer, my hope is that my time will be directed toward general
usability issues and on new parallelism implementation work.

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.


Miles T. Parker
Software Engineer
The Brookings Institution  1775 Mass. Ave. NW  Washington, DC  20036
mailto:address@hidden  voice 202.797.6136  fax 202.797.6319


                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.