Re: Laymans explaination for single level store?

[Steve Richter <stephenrichter@xxxxxxxxx>]

On Tue, 30 Nov 2004 13:53:01 -0500, cwilt@xxxxxxxxxxxx
<cwilt@xxxxxxxxxxxx> wrote:
> > -----Original Message-----
> 
> 
> > From: Walden H. Leverich [mailto:WaldenL@xxxxxxxxxxxxxxx]
> > Sent: Tuesday, November 30, 2004 1:29 PM
> > To: Midrange Systems Technical Discussion
> > Subject: RE: Laymans explaination for single level store?
> >
> >
> > >in order to allow programs
> > >to run on different hardware without re-compiling.
> >
> > I've always had a problem with this statement. Remove
> > observability and
> > how far do you get on hardware changes?
> 
> Actually, pretty far now-a-days.  Removing observability doesn't effect the
> system's ability to recreate the program object anymore.
> 
> So when we move to a 128bit CPU as long as the program was created recently
> (v4 or v5???) it won't matter if the program has observability or not.

I am not completely sure the 400 would be able to jump to a pointer
size > 64 bits.   There is a performance hit on the system when a
pointer is stored as 16 bytes rather than just storing the 8 byte
address part of the pointer. So I would not be surprised to learn that
a lot of the internal data structures of the system hold 8 byte
pointers instead of the full 16 bytes.  If so, then  increasing the
address size from 8 to 12 bytes might entail more than just a
recompile.

I think it is interesting that the 400 would benefit most from an
increase in the address size to 12 or 16 bytes  ( you could extend the
SLS to the entire network ).  But because of IBM's consolidation of
the PowerPC processor for all its platforms, that will not happen as
long as the 400 is a part of IBM.

-Steve

> 
> 
> 
> > Aren't you really recompiling
> > the program under the covers. Granted, observability doesn't store the
> > actual RPG source code (or cobol, or MI, or...) but it does store the
> > program template from which the "real" program is created, or
> > recreated,
> > no? The template is in effect the real source code, and the compilers
> > are simply translates from RPG, cobol, mi, etc. to the template.
> >
> > Now, I do give Frank and the others great respect for knowing
> > enough to
> > store the program template, and not rely on the "original"
> > source code.
> > But is it really fair to say you're not recompiling the code?
> >
> > -Walden
> >
> 
> I suppose one could make this argument, since "compiling" is generally
> accepted to mean converting to a binary format executable on the hardware.
> 
> On the other hand,  "compile" usually means from high-level source code.
> 
> So I think the answer is kinda, sorta, but not really ;-)
> 
> Charles
> 
> 
> --
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> or email: MIDRANGE-L-request@xxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/midrange-l.
> 
>