[Date Prev][Date Next][Subject Prev][Subject Next][ Date Index][ Subject Index]

Re: RE Which Computer



Paul Breeze wrote:
Patricia

Do the Asus boards allow you to use expanded (or is it extended, I can never remember) memory that
XY4 can utilise?

Thanks

Paul Breeze
This whole question is far more complex than it appears. It's a
complicated mix of hardware and opsys. Robert reports that (under W2K?)
on many mobos the USB drivers usurp the memory area that should be used
for EMS (expanded memory; that's the kind that Xy officially uses,
though I suspect that some XMS [extended] memory is also used, certainly
during "installation"--which is why you cannot "install" under 32-bit
Windows, but have to copy an existing setup). BUT I think that the
actual physical memory mapped as EMS by a DOS (or Win 3.1) app running
in a Virtual DOS Machine (and that's what's happening, even in 9x where
there is real DOS) can be anywhere in the system. It's not limited to
the 64K window, swapped in and out of HMA, that it was under DOS. (I've
proved this by having two DOS apps, Xy and dBase5, open; shelled to DOS
from both; and ran MEM /d. The SAME "memory addresses" were reported "in
use" by different apps. To my mind, that means that the "640K" of DOS is
just any block of 640K, plus any EMS, and Windows "tells" the DOS
session that it's the 640 K that it's expecting.)

What I need to do, and haven't had the time to, is carefully note down
the addresses Control Panel reports occupied and compare them with what
MEM reports--a project made more difficult by the fact that Control
Panel and MEM use a different scheme of writing memory addresses, and
I'm not fluent in translating between the two. What I can say is that I
do not find a lack of EMS memory available for Xy (regardless of what
Control Panel says) on any of my systems. Some are very old, and don't
have USB. But the three that are newer and do are all AMD CPUs. So
possibly there's something in the chipsets used with AMDs that preserves
that possibility. (Another reason, in addition to "Resist would-be
monopolists," to go with AMD chips.)

Here's a snippet of the report:
Xy and Dbase both in mem; Xy has focus, shelled to DOS, MEM reports:
 00800        240  (0K) COMMAND   Data
 0080F       5,728  (6K) COMMAND   Program
 00975       2,080  (2K) COMMAND   Environment
 009F7         80  (0K) COMMAND   Data
 009FC        304  (0K) EDITOR    Environment
 00A0F      116,528 (114K) EDITOR    Program
 02682        304  (0K) COMMAND   Data
 02695       5,536  (5K) COMMAND   Program
 027EF       2,080  (2K) COMMAND   Environment
 02871        336  (0K) MEM     Environment
 02886       90,464  (88K) MEM     Program
 03E9C      398,912 (390K) MSDOS    -- Free --
Xy and dBase both in mem; dBase has focus, shelled to DOS, MEM reports:
  00800  240 (0K)  DBASE  Environment
  0080F  51,696 (50K)  DBASE  Program
  014AE  2,064 (2K)  DBASE  Data
  0152F  480 (0K)  DBASE  Data
  0154D  1,040 (1K)  DBASE  Data
  0158E  8,208 (8K)  DBASE  Data
  0178F  64 (0K)  DBASE  Data
  01793  304 (0K)  DBASE  Data
  017A6  240 (0K)  COMMAND  Data
  017B5  256 (0K)  MEM  Environment
  017C5  144 (0K)  MSDOS  -- Free --

You can see that while some things keep out of each other's way, others
seem to be both in the same address.
By the by, on all systems, if I shell to DOS from Xy and type mem, free EMS memory is reported as 0--because Xy is using it, and the available amount is what is spec'd for Xy in its PIF.

Patricia M. Godfrey