Re: bytecode -> ssa -> bytecode

"Jim White" <jim@pagesmiths.com>
27 Jul 2000 21:36:07 -0400

          From comp.compilers

Related articles
bytecode -> ssa -> bytecode selby@ptibonum.scl.csd.uwo.ca (Jason Selby) (2000-07-23)
Re: bytecode -> ssa -> bytecode jim@pagesmiths.com (Jim White) (2000-07-27)
| List of all articles for this month |

From: "Jim White" <jim@pagesmiths.com>
Newsgroups: comp.lang.java.machine,comp.compilers
Date: 27 Jul 2000 21:36:07 -0400
Organization: @Home Network
References: 00-07-059
Keywords: optimize

"Jason Selby" <selby@ptibonum.scl.csd.uwo.ca> wrote in message
> My only thought is to ensure the class implements Cloneable and insert
> calls to the clone method to create the temps. Is this the only
> solution??


I'm quite certain no application programmer would want the JVM
(optimizing or otherwise) to be inserting its own clone calls.


I believe the answer is that you should just look for classes which are
known to be immutable (java.lang.String, Integer, etc). Optimizing
String would be the biggest single improvement you could ever hope for
with ordinary real world Java applications.


If this is some sort of really big win optimization, you could have a
switch where the programmer asserts that he believes all the Cloneable
classes in his system are well behaved (hard to know how he would know
though), in which case you could look for Cloneable as you suggest.


jim
------------------------------------------------------------
James P. White Netscape DevEdge Champion for IFC
IFC Exchange - Insanely great Java - http://www.ifcx.org


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.