Re: memory could not be "read"
- From: Paul <nospam@xxxxxxxxxx>
- Date: Sat, 11 Aug 2007 08:58:34 -0400
gremme wrote:
Hi<<snip>>
I got a problem with my computer.
now and then, sometimes very often, when I'm playing world of warcraft
a message pops up (looks like this but with different numbers from time to time:
error#132 (0x85100084) fatal exception
program: c:\program\world of warcraft\wow.exe
exception: 0xc0000005 (access_violation) @ 001b:4d67375
the instruction at "0x4d657375" referenced memory at "0x4d657375".
the memory could not be "read".
then the game will shut down after I press ok.
Does anyone know what might cause this? I'll be very grateful for any help.
Below I've included system specs and an errorlog
Here's a summary of my system specs from the directX diagnostics:
OS: Microsoft windows XP professional(5.1, build 2600)
Systemmanufacturer: NVIDIA
Systemmodel: AWRDACPI
BIOS: Phoenix - AwardBIOS v6.00PG
Processor: AMD Athlon(tm) 64 X2 Dual Core Processor 4200+, MMX, 3DNow(2 CPUs), Memory: 2048MB RAM
Exchangefile: 711 MB used, 3228 available DirectX-version: DirectX 9.0c(4.09.0000.0904)
Here's an errorlog from world of warcrafts errorlogs:
There is an example here, where a guy noticed his WOW problems went away
when he fixed his RAM timings. I found about 900 references to problems
like this - in many cases, the users couldn't figure out what was wrong.
(Note - sometimes problems like this are a coding problem with the game,
and a patch can fix it. Finding a WOW forum might have more details about
the patching history of your game.)
http://www.houseofhelp.com/forums/showthread.php?p=281455
To test the stability of a system, your options are to try the "Torture Test"
menu option of Prime95 (mersenne.org) or use Orthos. Orthos is basically
Prime95, only runs two copies on a dual core processor for better testing.
http://sp2004.fre3.com/beta/orthos_exe_20060420.cab (file to download)
http://sp2004.fre3.com/beta/beta2.htm (the download page)
http://sp2004.fre3.com/ (Stress Prime 2004 main page)
I believe I've seen yet another version of a Prime95 test tool, that
supports quad core testing, but I haven't located details on it yet.
I've seen a picture of the screen, but no download link.
Memtest86+ and Prime95 testing are complementary. Memtest86+ can test
all of a systems memory, and excels at "stuck at" bit problems (i.e.
permanently bad memory). Prime95, on the other hand, can only test the
free memory in the system, so if you are lucky, maybe 75% of the memory
gets tested. But Prime95 is a more sensitive test, and finds errors
even when memtest86+ runs clean for hours on end. So both tests must
be used, to get a good feeling about the RAM.
I have had memtest86+ detect a "bad spot" (consistent errors) on a piece
of memory, so the program does work and it is worth running. It just
doesn't tell you everything worth knowing.
Also, for dual core systems and gaming, some games don't like to run on
a dual core. For those, start the game, then alt-tab out and bring up
the task manager (ctrl-alt-delete). Find the program in the list of
processes, then set the affinity to a single core. That prevents the
program from being moved from core to core by the scheduler. On some
programs, that prevents problems and makes the game run smoother. It
is only worth trying, if some Googling turns up a lot of reports that
affinity is necessary to get a particular game to run - someone should
already have tried the experiment, to make it worthwhile for you to
use the hack.
Another way to control affinity (i.e. make program stick to using one
core), is use a program like this, and prepare a shortcut to launch
the program. There are other programs besides this one.
http://www.activeplus.com/us/freeware/runfirst/
I would concentrate on testing the memory first, because a stable
system is more important than being able to run one game. If you
cannot run something like Orthos stable, then errors could crop
up in any program you use.
Paul
.
- References:
- memory could not be "read"
- From: gremme
- memory could not be "read"
- Prev by Date: Re: memory could not be "read"
- Next by Date: Re: Hard Drive Space Disappeared
- Previous by thread: Re: memory could not be "read"
- Index(es):
Relevant Pages
|