Re: Win 98 2E Memory Leaks

[Allen Schaaf <soundbyte@xxxxxxxxxxxxxxxxxxx>]

Dan,

 From what I can tell you are doing the right things, but others probably 
can add more.

My points are that I would add more physical memory - it seems to do more 
good than virtual memory by a whole bunch.

Second, I don't trust Norton.  I've had too many conflicts with it and 
switched a couple of years ago to System Suite from Ontrack. Better 
integrated and easier to use.  It also doesn't seem to leak memory like 
Norton used to.  I'm not versed in the current Norton but the people where 
I'm working keep having problems, I give them a demo of System Suite and 
the problems seem to go away.

Third, I'd be wary of programs that claim to cure memory leaks.  My 
experience is that memory leaks are internal to the programs themselves and 
need to be solved by the original programmers.  But it is possible to 
re-claim memory that is sitting idle and forgotten about by the program, I 
just don't know if they work well.

As to IE, go to www.fuckmicrosoft.com, read and heed the article on the 
"Hidden Files".  My 98SE machine is much more stable now.

As always, YMMV.


Allen Schaaf
Sr. Tech Writer
Fourelle Systems, Inc.

Who says bad manuals aren't a risk to your life?  Just ask the passengers 
of the jet where the engine caught fire because the company's maintenance 
manual was wrong about how to install one key bolt.  (NTSB Report on GE CF6 
engine fire, American Airlines flight 574, July 9, 1998. 
<http://www.ntsb.gov/publictn/1999/AAB9903.htm>)





At 11:32 AM 3/3/02, Dan Emory wrote:
>I have a Win 98 2E platform with 128 MB of memory. Virtual Memory is set 
>to a minimum size of 1 GB. My main disk is very close to 100% optimized. 
>Using Norton System Doctor sensors, I can monitor memory, virtual memory, 
>and swap file usage, and current date/time. As memory usage approaches 
>100% with all programs inactive, I initially note that the time begins 
>lagging behind real time, and computer response slows down appreciably. 
>Continued operation under this condition almost inevitably produces 
>anomalies, requiring a  shutdown or restart. It's also interesting to note 
>that post-restart/shutdown memory usage is 2% higher after a 
>shutdown/restart than it is after a shutdown/power-off/power-on sequence.
>
>I've been using the System Doctor Memory Usage sensor to determine memory 
>usage before, during, and after opening, using, and closing various 
>programs. Virtual Memory and Swap File Usages remained unchanged during 
>these tests. But the difference between the Before and After memory usage 
>values can be significant, and indicate the magnitude of the memory leak 
>attributable to each program.
>
>I've found the main culprits to be Eudora, Internet Explorer, and Acrobat 
>4.05 Distiller. Surprisingly, FrameMaker+SGML 5.5.6 and 6.0 seem to 
>produce tolerable memory leaks, although, in some cases, the size of the 
>leak becomes more severe.
>
>Questions:
>
>1. Does my methodology for detecting the amount of memory leakage produced 
>by various programs appear to be valid?
>
>2. Have any of you used a similar methodology to identify the programs 
>which are the main culprits, and if so, do your findings agree with mine?
>
>3. I've looked on the internet for programs capable of eliminating memory 
>leaks, but I'm still leery of installing any of them. Does anyone have any 
>recommendations about which program works best on a Win 98 2E platform?
>


** To unsubscribe, send a message to majordomo@omsys.com **
** with "unsubscribe framers" (no quotes) in the body.   **