News said:
Is this a different Wintop than the one put out by Microsoft? That
one shows memory usage only on a selected application basis, though it
does show CPU usage.
It is Microsoft's.
Depending on what you're doing this info may also help:
"Memory Hogs Identified
How to assess a program's memory usage
by Alan Martin
Unless it is done methodically, memory measurements can be misleading.
This article describes how to compare the memory used by competing
programs, such as a shortlist of web browsers or a few different shells.
That's an important factor when choosing between competing programs if
you have 32 MB of RAM. If you have 512 MB, you can ignore all this and
choose according to their features rather than their memory usage.
There's also a method to test for "dirty" apps which fail to release all
their allocated memory when they are closed.
It's very hard to judge memory usage just from the exe file size, because:
* Some exes are compressed and expand when in RAM.
* They may demand more RAM beyond the exe size, for data.
* Some use more system dlls than others. System dlls (i.e. those
from Microsoft, not from the application maker) are no burden if they
are the most common ones, because they are probably already loaded for
another app anyway. But even the smallest Visual Basic program, for
example, may have a significant impact on free RAM if it is the only VB
program.
As a result, one can only *guess* a program's RAM usage by doing sums on
its exe size and dlls. Which of the dlls that a specific program calls
should be counted against it? It depends! (spot the pun)
This pragmatic approach is far more reliable: Use a memory meter, such
as Infoman2 from
http://www.systemsbysteve.com/mypicks.htm which shows
free physical (not virtual) RAM. Note the free RAM before and after
launching the app. But please note:
When testing memory loads, you will need to set the min and max cache
values equal (not a good permanent setting) else with an elastic cache
the memory comparisons will be meaningless. In system.ini under [386enh]
type two lines:
mincachesize=8192
maxcachesize=8192
and then reboot so it takes effect. Those numbers are KB so you will get
an 8 MB cache. It doesn't matter as long as both values are equal. Don't
forget to delete or ;rem those new lines when you've finished comparing
the programs.
Then test each contender immediately after a reboot *and* after opening
a few of your often used apps, such as Windows Explorer (or your
favourite FM) and notepad. Each time, make the process identical, from
reboot until the app to be tested is launched. Then note how much free
RAM it leaves. That's a time consuming process, so write down the
results! I don't know of a quicker way which will avoid all the
uncertainties. Of course an individual result has little meaning. It's
the difference between the apps' results which matters.
Oddly, for this purpose, a more sophisticated memory monitor is no more
useful than a simple one. They may offer to add up the exe size in ram
and the dll dependencies but, as explained above, that can be misleading.
With a similar method you can also discover dirty apps which don't
release all their RAM on closing. That's more of a problem with
Win95/98/ME than with NT/2K/XP because the latter have better memory
management.
With a dirty app, if you use and then close it several times, the free
RAM gradually declines. Don't just open and close it; open and edit a
real document file each time. If it's a browser, browse with it.
This time, you are not comparing between apps; you are comparing after
each of several runs of the same app. Also with the first job, described
above, you note the RAM while each app is open. With this job, you note
it after each time the one app is closed.
With some versions of Internet Explorer under Win98, I found that after
surfing the www for a while, the PC would bog down entirely and would
need a reboot. The system remained sluggish even after IE was closed.
One didn't even need a memory meter to notice how "dirty" that
application was!
Alan Martin"
Cheers!
Son Of Spy
--
Some You Won't Find Anywhere Else...
http://www.sover.net/~wysiwygx/index.html
. --- . . - - - - - - - - - - - -
/ SOS \ __ / Freeware - - - - - -
/ / \ ( ) / - - - - -
/ / / / / / / \/ \ - - - -
/ / / / / / / : : - - -
/ / / / / ' ' - -
/ / //..\\
=====UU==UU=====
'///||\\\'
' '' '
