>Ok I hacked together a VM via CGeneratorEnhancements-ajh.cs with a 
>bit of hand coding.
>
>so...
>
>on a 68k BSD box with GCC the new numbers are
>
>1,614,205 bytecodes/sec and 57,652 sends/sec
>
>versus my previous one using the jumptable modification
>
>1,550,387 bytecodes/sec and 55,080 sends/sec
>
>versus what I started with
>
>1,439,884 bytecodes/sec and 51,098 sends/sec
>

Just out of curiosity I tried this with a VM at 3.0.22 (this actually 
builds!) for 68K macs under os 7.55 - os9.2.x. Interestingly enough I 
found that the brain dead optimizer just assigns registers to the 
most use variables. Thus t1-tx are targeted. On applying 
CGeneratorEnhancements-ajh.cs we end up with a zillion locals within 
the routine, then the optimizer just gives up and mostly produces 
code that requires you to move items to and from the stack.

However this got me thinking. I'm wondering if someone wants to build 
a version of GCC for 68K for MPW, much like the PPC one we've got for 
Jitter 3. (only so many hours in a day, and I'd like to think there 
is someone out there who's idle and really really needs a faster 68K 
VM).


-- 
--
===========================================================================
John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
===========================================================================