Kim Timothy J (tjk1846@ucs.usl.edu) wrote:
: Does anybody have the new Borland C++ 32 bit Dos Extender (Power Pack)?
: They say that all you need to do is flip a switch on your compiler and
: voila - 4 megs at 32 bits (or 16 at 16 bits). Sounds very cool (except
: that you still have to develope in the Windows IDE).

I have a question :

4 megs at 32 bits - should it not be 4 giga at 32 bits
('nobody' have that much but still 8-) )

16 at 16 bits - that is then 16 megs at 16 bits

Or does the 32 bit version realy only give you 4 megs
to play with ? 8-(

Thanks for any enlightment,

Hans

In article <Cr6FtL.HGx@dcs.gla.ac.uk>, hans@dcs.gla.ac.uk (Hans
Stadtler) writes:
|>
|> I have a question :
|>
|> 4 megs at 32 bits - should it not be 4 giga at 32 bits
|> ('nobody' have that much but still 8-) )
|>
|> 16 at 16 bits - that is then 16 megs at 16 bits
|>
|> Or does the 32 bit version realy only give you 4 megs
|> to play with ? 8-(
|>
|> Thanks for any enlightment,
|>
|> Hans
|>

The 16-bit version uses the 80286 protected mode, which allows the
processor to access all 24 bits of its address space: 2^24 = 16M.
It is more or less an accident that the 286 is a 16-bit processor
and has a 16MB address space.  After all, the 8086 is a 16-bit
processor as well, but it only has a 1MB address space.

This is because the 8086 has 20 address lines.  The 80286 has
24 address lines.  Only 20 of these are used in real mode of course.

Now, the 80386 is a 32 bit processor, and has 32 address lines.
The difference here is that it supports virtual memory.  In protected
mode, memory is described by segments.  These segments are the basis
for the protection model.  For example, you wouldn't be able to write
to your code segment.  That allows the code segments to be shared.
The segmentation mechanism translates a segment/offset pair into a
linear address, which can be used as the physical address (32 bits).
Or, if paging is enabled, the linear address can be paged, which
allows the system to have 16-terabytes (64 bits).  Only 4gig (32 bits)
of this can be "present" at any given time.  The rest must be paged out.
This is because the 386 only has the 32 address lines.

The 32-bit PowerPack mode will use VM if provided (by Windows for example)
and that would allow you to use quite a bit of memory.  But a single
segment can only be 4GB, so no one program would likely be close to 16TB.
What this does mean is that if you have, say 8MB of RAM, then your DOS
programs will have access to _more_ than that, say 12MB, under Windows.
Not to mention the potential memory savings of the DLL feature...

One more comment:
  386-SX / 486-SX machines typically have only 24 address lines.

Well, I bought into it when I was calling for the CD with patches. I looked at
the press release and tech info on the ftp server, and was impressed with the
claimed DLLs for DOS and Windows (share one DLL for both environments).

However it isn't available yet! You can give them your CC # but it most likely
won't arrive at your door step till July (that's what the rep told me). They
won't bill you till the box is shipped.

In article <2t2vjq$d2l@rouge.usl.edu> tjk1846@ucs.usl.edu (Kim Timothy J)
writes:>From: tjk1846@ucs.usl.edu (Kim Timothy J)
>Subject: New BC++ 4.0 DOS Extender
>Date: 7 Jun 1994 23:24:42 GMT

>Does anybody have the new Borland C++ 32 bit Dos Extender (Power Pack)?
>They say that all you need to do is flip a switch on your compiler and
>voila - 4 megs at 32 bits (or 16 at 16 bits). Sounds very cool (except
>that you still have to develope in the Windows IDE).

>Should I make another investment? Is it that much easier than Pharlap
>DOS Extender? If what they state in their 'comic book' upgrade mailer
>is true...I must have it ;-).

>So whose got it already and has used it enough to comment.

>Tim "should I wait for the next upgrade" Kim







Jeff Macdonald
Ziff Information Services
Ziff Communications
10 President's Landing
Medford, MA
jeff@zis.ziff.com

Kim Timothy J (tjk1846@ucs.usl.edu) wrote:
: Does anybody have the new Borland C++ 32 bit Dos Extender (Power Pack)?
: They say that all you need to do is flip a switch on your compiler and
: voila - 4 megs at 32 bits (or 16 at 16 bits). Sounds very cool (except
: that you still have to develope in the Windows IDE).

I too am interested in this product.  According to the sales rep at Borland,
they won't be shipping this product until Jun 17.  My order for one is
already in...

: Should I make another investment? Is it that much easier than Pharlap
: DOS Extender? If what they state in their 'comic book' upgrade mailer
: is true...I must have it ;-).

I use Pharlaps DOS extender now (as well as TurboVision), and intend to test
this new product against it.  Depending upon the results, I may be able to
forego Pharlaps $1500.00 royalty payment for unlimited distribution of
products based upon their 286-Extender...

: So whose got it already and has used it enough to comment.

Sit tight, and you should know before the end of June...

: Tim "should I wait for the next upgrade" Kim

Fred Hensley
TradeWinds Software, Inc.
EMAIL: fhensley@eskimo.com
--
*****************************************************************************
* "No man can be condemned for owning a dog.  As long * Fred Hensley        *
*  as he has a dog, he has a friend; and the poorer   * fhensley@eskimo.com *
*  he gets, the better friend he has." - Will Rogers  * Thank God for pets! *
*****************************************************************************

Does anybody have the new Borland C++ 32 bit Dos Extender (Power Pack)?
They say that all you need to do is flip a switch on your compiler and
voila - 4 megs at 32 bits (or 16 at 16 bits). Sounds very cool (except
that you still have to develope in the Windows IDE).

Should I make another investment? Is it that much easier than Pharlap
DOS Extender? If what they state in their 'comic book' upgrade mailer
is true...I must have it ;-).

So whose got it already and has used it enough to comment.

Tim "should I wait for the next upgrade" Kim