From: MooseFET on
On May 11, 9:24 pm, John Larkin
<jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote:
> On Tue, 11 May 2010 20:12:47 -0700,
>
>
>
> "JosephKK"<quiettechb...(a)yahoo.com> wrote:
> >On Tue, 11 May 2010 06:24:27 -0700, John Larkin
> ><jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote:
>
> >>On Mon, 10 May 2010 21:14:10 -0500, "Tim Williams"
> >><tmoran...(a)charter.net> wrote:
>
> >>>"John Larkin" <jjlar...(a)highNOTlandTHIStechnologyPART.com> wrote in message
> >>>news:hjahu5d35h5ej11fe9ml9hlvepjceek3jn(a)4ax.com...
> >>>> Sure, but why do they make you install a virtual machine and a whole
> >>>> nother OS to run 16 bit apps? Why didn't they include a dosboxy thing?
>
> >>>John,
>
> >>>It's not a Windows thing.  In 64 bit ("long") mode, 16 bit instructions
> >>>don't work (either generate a fault or are mapped to other instructions).
> >>>32 bit is the new 16 bit (32 bit programs run in long mode for
> >>>compatability).
>
> >>Of course it's a Windows thing. Surely Microsoft has a programmer good
> >>enough to write a dosboxy thing that will run 16-bit code. Or they
> >>could hire one. They just chose not to.
>
> >>>Anything that emulates (NOT virtualizes) 16 bit operation is an additional
> >>>program, like DOSBox or the Virtual Machine stuff.
>
> >>Right. Win7/64 doesn't include that. Which means I'm in no hurry to
> >>use it.
>
> >>John
>
> >Switching between 64 bit mode and other modes is really quite expensive
> >in CPU cycles, cache misses, and more.  Doing just a little of that will
> >eat over 10% of the processor and memory bandwidth.  I really is not all
> >that compatible if you are switching modes all the time.
>
> Why would it matter that a program that ran fine on a 40 MHz 486 uses
> 10% of one core of a 2 GHz quad-core CPU? If it were rewritten into
> modern bloatware, it might use 60%!
>
> I have a lot of 16-bit engineering apps I'd still like to run.

It was the desire to run my DOS programs without constantly having the
OS butt in a shut them down, have them able to do hardware actions on
the parallel port, and have my DOS programs access USB serial ports
that caused me to decide to select a 32 bit OS after I had been
running a 64 bit version for a while.

The virtual machines running on the 64 bit never seemed to be able to
do the hardware related things with anything like reasonable timing if
they could manage it at all.

I still use DOS tools for compiling microcontroller code. Running on
the 2.2GHz machine it takes about 1 hour to get the full set of
files. Back in the days of the PC-XT, it would have taken a little
more than over night. There are about 20 products all based on the
same code so if I make a change, I often have to recompile to whole
set.

BTW: Each of them is about 300K bytes to give a scale for the job.


From: Joel Koltner on
"MooseFET" <kensmith(a)rahul.net> wrote in message
news:f68a4980-53ff-4d32-aea1-d17b7d6e2cc2(a)r21g2000prr.googlegroups.com...
> I still use DOS tools for compiling microcontroller code. Running on
> the 2.2GHz machine it takes about 1 hour to get the full set of
> files.

Are they all just "command line"-driven (i.e., no GUIs)? If so, if you could
gain access to the original source code, there's a good chance it'd just be an
"easy" re-compile and you'd get ever *more* performance... :-)


From: John Larkin on
On Wed, 12 May 2010 08:48:06 -0700, "Joel Koltner"
<zapwireDASHgroups(a)yahoo.com> wrote:

>"MooseFET" <kensmith(a)rahul.net> wrote in message
>news:f68a4980-53ff-4d32-aea1-d17b7d6e2cc2(a)r21g2000prr.googlegroups.com...
>> I still use DOS tools for compiling microcontroller code. Running on
>> the 2.2GHz machine it takes about 1 hour to get the full set of
>> files.
>
>Are they all just "command line"-driven (i.e., no GUIs)? If so, if you could
>gain access to the original source code, there's a good chance it'd just be an
>"easy" re-compile and you'd get ever *more* performance... :-)
>

Some are command-line, some are console-type windows with menus, some
have DOS-type graphics. Some are my code with sources available, some
aren't.

It's a lot "easier" to not bother with Win7.

Why doesn't Win7 just run everything?

Answer: because the judge that decided to break up Microsoft was
overruled.

John


From: MooseFET on
On May 12, 8:48 am, "Joel Koltner" <zapwireDASHgro...(a)yahoo.com>
wrote:
> "MooseFET" <kensm...(a)rahul.net> wrote in message
>
> news:f68a4980-53ff-4d32-aea1-d17b7d6e2cc2(a)r21g2000prr.googlegroups.com...
>
> > I still use DOS tools for compiling microcontroller code.  Running on
> > the 2.2GHz machine it takes about 1 hour to get the full set of
> > files.
>
> Are they all just "command line"-driven (i.e., no GUIs)?  If so, if you could
> gain access to the original source code, there's a good chance it'd just be an
> "easy" re-compile and you'd get ever *more* performance... :-)

On some of the command line tools, I do have source code.
Unfortunately, it is in 16 bit C and I know from experience
that trying to find all the places where the code assumes
an "int" is exactly 2 bytes and a pointer is exactly 4 would
take a while.

I did not write the code in question but I have had to maintain
it and fix bugs. It is a bit of a "pigs breakfast"internally.

The full recompile is rare enough that I am not strongly
motivated to start trying to port it over to native 32 bit
linux. Besides, there are parts that I don't have the source
for.
From: JosephKK on
On Tue, 11 May 2010 18:48:17 -0700, Copacetic
<Copacetic(a)iseverythingalright.org> wrote:

>On Tue, 11 May 2010 03:30:14 -0700, "JosephKK"<quiettechblue(a)yahoo.com>
>wrote:
>
>> back then it was to get 32 bit softs you had to
>>give up 8 bit softs.
>
> More bullshit.
>
> My MFM drives worked fine on a 486 as long as it had the slot for it.
>
> It also booted. Guess what was loaded on it.

And just what does the then already obsolete hardware have to do with the
design of the software?