VW 2.0 Decompiler optimize switch 
Author Message
 VW 2.0 Decompiler optimize switch

According to the Decompiler class comment ". . .we have choosen to provide
the decompiler with a switch that controls whether it will produce code
that can be decompiled . . .".  I've looked in the Compiler classes but I
can't seem to find a switch or mention of it.  Does anyone know what the
switch is (ie., method/inst var/class var/etc.) and how to turn it on or
off?  Thanks for any help you may offer.

--
Rames Creel

San Joaquin Delta College



Tue, 16 Dec 1997 03:00:00 GMT  
 VW 2.0 Decompiler optimize switch

Quote:

>According to the Decompiler class comment ". . .we have choosen to provide
>the decompiler with a switch that controls whether it will produce code
>that can be decompiled . . .".  I've looked in the Compiler classes but I
>can't seem to find a switch or mention of it.  Does anyone know what the
>switch is (ie., method/inst var/class var/etc.) and how to turn it on or
>off?  Thanks for any help you may offer.

See the 'mustBeSimple' inst var of ByteCodeStream.

___________________________________________________________________
Gordon Mohr                Talk Small and carry a big class library



Fri, 19 Dec 1997 03:00:00 GMT  
 VW 2.0 Decompiler optimize switch


: >According to the Decompiler class comment ". . .we have choosen to provide
: >the decompiler with a switch that controls whether it will produce code
: >that can be decompiled . . .".  I've looked in the Compiler classes but I
: >can't seem to find a switch or mention of it.  Does anyone know what the
: >switch is (ie., method/inst var/class var/etc.) and how to turn it on or
: >off?  Thanks for any help you may offer.

: See the 'mustBeSimple' inst var of ByteCodeStream.

What sort of performance improvement might I see if I turned this flag off?

I guess it's probably architecture-specific, but what is the best figure
I could hope for?

Is the warranty void if I touch it :-) ?

Neil McKee
Cambridge Quality Management Ltd.



Sun, 21 Dec 1997 03:00:00 GMT  
 VW 2.0 Decompiler optimize switch

Quote:


>: See the 'mustBeSimple' inst var of ByteCodeStream.

>What sort of performance improvement might I see if I turned this flag off?

>I guess it's probably architecture-specific, but what is the best figure
>I could hope for?

Sorry -- I don't have any figures for that. You'd have to run your own
tests. A possible test:

  - Run the Benchmark suite contained in the VisualWorks Advanced Tools
    product in a normal image. Record the results.
  - Change the 'mustBeSimple' default in ByteCodeStream>>initialize
  - Recompile every method in the image. (Something like the following
    might work: "Smalltalk recompileMethodsForWhich: [ :meth | true]"
    However, it might not -- I'm not sure how currently active contexts
    will handle this total recompile.) This will take a while.
  - Run the Benchmark suite again, comparing the results to the first
    run.

___________________________________________________________________
Gordon Mohr                Talk Small and carry a big class library



Tue, 23 Dec 1997 03:00:00 GMT  
 
 [ 4 post ] 

 Relevant Pages 

1. VW 1.0 to VW 2.0 Upgrade Issues

2. Use of VW 2.0 Notebook widget in VW 1.0

3. OLE 2.0 Support for VW 2.0

4. VW Switching between canvas window specs

5. VW 1.0/ 2.0 + Apple Events ?? (=> Mosaic + VW 1.0/2.0)

6. Optimizing programmer vs optimizing compiler (was: Re: Assembly vs c/c++)

7. Image Manipulation VW 2.0

8. Curve-fitting objects for PP VW 2.0 Image

9. Fast grab of Screen in PP VW 2.0

10. Image Manipulation VW 2.0

11. Help: Headless image in VW 2.0

 

 
Powered by phpBB® Forum Software