[JSR308] Java compilers that perform common subexpression elimination

Neal Gafter gafter at google.com
Sat May 19 00:53:18 EDT 2007

I don't know of any java-to-bytecode compilers that perform optimizing
transformations. The only thing javac does even remotely like this is that
javac does not generate bytecode for dead code (for example code in the
contolled statement of an if(false) generates nothing in the bytecode).

On May 18, 2007 8:57 PM, Michael Ernst <mernst at csail.mit.edu> wrote:
> If a compiler performs common subexpression elimination on two
> (incompatibly-annotated) expressions or statements, then there is no good
> choice for how to annotate the generated classfile.  I am not sure whether
> this is a purely hypothetical problem or a real one.
> As a first step in determining this, can someone, or some group of people,
> volunteer to create a catalog of widely used Java-to-bytecode compilers
> and
> whether they perform this or other problematic optimizations?  (A list of
> other problematic optimizations would also be helpful.)  Even if you can
> only report the results for one compiler, that would be helpful.
>                    -Mike
> _______________________________________________
> JSR308 mailing list
> JSR308 at lists.csail.mit.edu
> https://lists.csail.mit.edu/mailman/listinfo/jsr308
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.csail.mit.edu/pipermail/jsr308/attachments/20070518/98dc0d20/attachment.html

More information about the JSR308 mailing list