help me interpret MemTest86 results?

Hi folks,

I've got an abit KT7(A?) setup running WinME that I just bought some
new RAM for. I got 3 new sticks of generic PC133 512 Meg, for a total
of 1.5 Gig. I tracked down some initial issues to system.ini settings
that allow WinME to live with ~1 Gig RAM, but the system is still
unstable with lots of recoverable blue screens and random 1-second
screen blanks. Just to see, I made a MemTest86 (v3.2) boot disk and
ran for 15+ hours on the auto settings. My results are below.
Unfortunately I don't know what to do with this info. Is there
anything useful I can glean from it other than that there's at least
one bad spot? Any way to tell which of the 3 sticks it is, for
instance, without running MemTest on them one at a time? Any help
appreciated.

MemTest-86 v 3.2 : Pass [...]
AthlonXP (0.13) 1733Mhz : Test [...]
L1 Cache : 128 10633MB/s : Test [...]
L2 Cache : 256K 3385MB/s : Testing [...]
Memory : 1536M 364MB/s : Pattern [...]
Chipset : VIA KT133/KT133A

WallTime Cached RsvdMem MemMap Cache ECC Test Pass Errors
17:02:xx 1536M 76K e820-Std on off Std 8 2

Ecc Errs
0
__________________________________________________ _________________________
Tst Pass Failing Address Good Bad Err-Bits Count Chan
2 0 000320b4e38 - 800.2MB ffffffff fffffeff 000001 2 1


--David

> Hi folks,
>
> I've got an abit KT7(A?) setup running WinME that I just bought some
> new RAM for. I got 3 new sticks of generic PC133 512 Meg, for a total
> of 1.5 Gig. I tracked down some initial issues to system.ini settings
> that allow WinME to live with ~1 Gig RAM, but the system is still
> unstable with lots of recoverable blue screens and random 1-second
> screen blanks. Just to see, I made a MemTest86 (v3.2) boot disk and
> ran for 15+ hours on the auto settings. My results are below.
> Unfortunately I don't know what to do with this info. Is there
> anything useful I can glean from it other than that there's at least
> one bad spot? Any way to tell which of the 3 sticks it is, for
> instance, without running MemTest on them one at a time? Any help
> appreciated.


*any* error at all means the RAM has failed...
you should try one stick at a time to see if just one perhaps is
defective...
but in general...stay away from generic RAM

On 22 Aug 2005 22:21:19 -0700, "David"
<david.lindsey@mindspring.com> wrote:

>Hi folks,
>
>I've got an abit KT7(A?) setup running WinME that I just bought some
>new RAM for. I got 3 new sticks of generic PC133 512 Meg, for a total
>of 1.5 Gig. I tracked down some initial issues to system.ini settings
>that allow WinME to live with ~1 Gig RAM, but the system is still
>unstable with lots of recoverable blue screens and random 1-second
>screen blanks.
>
> Just to see, I made a MemTest86 (v3.2) boot disk and
>ran for 15+ hours on the auto settings. My results are below.
>Unfortunately I don't know what to do with this info. Is there
>anything useful I can glean from it other than that there's at least
>one bad spot? Any way to tell which of the 3 sticks it is, for
>instance, without running MemTest on them one at a time? Any help
>appreciated.
>
>MemTest-86 v 3.2 : Pass [...]
>AthlonXP (0.13) 1733Mhz : Test [...]
>L1 Cache : 128 10633MB/s : Test [...]
>L2 Cache : 256K 3385MB/s : Testing [...]
>Memory : 1536M 364MB/s : Pattern [...]
>Chipset : VIA KT133/KT133A
>
>WallTime Cached RsvdMem MemMap Cache ECC Test Pass Errors
>17:02:xx 1536M 76K e820-Std on off Std 8 2
>
>Ecc Errs
> 0
>_________________________________________________ __________________________
>Tst Pass Failing Address Good Bad Err-Bits Count Chan
> 2 0 000320b4e38 - 800.2MB ffffffff fffffeff 000001 2 1
>
>
>--David

After 15 hours, after 8 passes, it would have produced many
many errors if the memory was physically damaged. This
looks more like a (barely) marginal, almost stable
configuration but untimately you need to either replace the
memory or try adjusting some bios settings to more
conservative memory timings, like raising the CAS if
possible. That WIndows had the errors is more evidence of
it.

If you had system your system.ini to a 512MB limit, try a
lower value like 384MB. From above memtest results the
error rate seems lower than would be expected to cause the
degree of problem you report in WinME- so I suspect ME might
have remaining issues using 1GB... BUT there's no point
worrying about that till memtest shows zero errors.

Consider that it took 15 hours to find a mere 2 errors. You
might need to retest (after making changes) for a full day
to have any reassurance, since a slower memory bus with more
memory just takes longer to test.

When you say that the system is now unstable, I think you've just showed
that your "fix" for getting WinMe to use more than 512 MB of RAM is wrong.

--
DaveW



"David" <david.lindsey@mindspring.com> wrote in message
news:1124774479.778770.209390@g43g2000cwa.googlegr oups.com...
> Hi folks,
>
> I've got an abit KT7(A?) setup running WinME that I just bought some
> new RAM for. I got 3 new sticks of generic PC133 512 Meg, for a total
> of 1.5 Gig. I tracked down some initial issues to system.ini settings
> that allow WinME to live with ~1 Gig RAM, but the system is still
> unstable with lots of recoverable blue screens and random 1-second
> screen blanks. Just to see, I made a MemTest86 (v3.2) boot disk and
> ran for 15+ hours on the auto settings. My results are below.
> Unfortunately I don't know what to do with this info. Is there
> anything useful I can glean from it other than that there's at least
> one bad spot? Any way to tell which of the 3 sticks it is, for
> instance, without running MemTest on them one at a time? Any help
> appreciated.
>
> MemTest-86 v 3.2 : Pass [...]
> AthlonXP (0.13) 1733Mhz : Test [...]
> L1 Cache : 128 10633MB/s : Test [...]
> L2 Cache : 256K 3385MB/s : Testing [...]
> Memory : 1536M 364MB/s : Pattern [...]
> Chipset : VIA KT133/KT133A
>
> WallTime Cached RsvdMem MemMap Cache ECC Test Pass Errors
> 17:02:xx 1536M 76K e820-Std on off Std 8 2
>
> Ecc Errs
> 0
> __________________________________________________ _________________________
> Tst Pass Failing Address Good Bad Err-Bits Count Chan
> 2 0 000320b4e38 - 800.2MB ffffffff fffffeff 000001 2 1
>
>
> --David
>

"DaveW" <none@zero.org> wrote in message
news:MKKdnZ2dnZ1KxYH3nZ2dnWM2lt6dnZ2dRVn-zZ2dnZ0@comcast.com...
> When you say that the system is now unstable, I think you've just showed
> that your "fix" for getting WinMe to use more than 512 MB of RAM is wrong.

Simple solution, get rid of WinME.

Michael

Try Gold Memory, from www.goldmemory.cz , because it usually finds
errors faster and takes roughly 40 minutes to test a 512MB module.
www.realworldtech.com rated it better than even MemTest86, and while
it's found errors missed by MemTest86 in some of my modules, MemTest86
also found errors that Gold Memory missed.

Some mobos just aren't reliable with more than two memory modules, at
last not at their standard settings. OTOH lots of modules I've tried
that had no-name chips on them failed testing, even major brand
modules, like Kingston.

Solid suggestions all around. Thanks very much. The first thing I
have tried is setting the Delay DRAM Read Latch at 0.5ns instead of
Auto through SoftMenuIII and I haven't had a blue screen in a full day.
We'll see if that or other BIOS fine tuning does the trick. Fingers
crossed.

DaveW wrote:
> When you say that the system is now unstable, I think you've just showed
> that your "fix" for getting WinMe to use more than 512 MB of RAM is wrong.

I added the "MaxPhysPage=3C000" statement to 386Enh and the
"MaxFileCache=523264" statement to system.ini. Are you aware of any
other specific things that would help? The party line from MS seems to
be that WinME will not support more than 1 Gig. I was under the
impression that meant ranges between 512 and 1024 Meg could be made
stable.

Cheers.
--David

"David" <david.lindsey@mindspring.com> wrote in message
news:1124934582.801196.170320@o13g2000cwo.googlegr oups.com
> Solid suggestions all around. Thanks very much. The
> first thing I have tried is setting the Delay DRAM Read
> Latch at 0.5ns instead of Auto through SoftMenuIII and I
> haven't had a blue screen in a full day. We'll see if
> that or other BIOS fine tuning does the trick. Fingers
> crossed.
>
> DaveW wrote:
> > When you say that the system is now unstable, I think
> > you've just showed that your "fix" for getting WinMe to
> > use more than 512 MB of RAM is wrong.
>
> I added the "MaxPhysPage=3C000" statement to 386Enh and
> the "MaxFileCache=523264" statement to system.ini. Are
> you aware of any other specific things that would help?
> The party line from MS seems to be that WinME will not
> support more than 1 Gig. I was under the impression that
> meant ranges between 512 and 1024 Meg could be made
> stable.
>
> Cheers.
> --David

on memtest, only 0 error's will do
you generaly only need to run 1 pass of memtest, run a few pass's to be sure
if you like

winme, oh get rid of it and get XP, we didn't call it muppet edition for
nothing...

On 24 Aug 2005 18:49:42 -0700, "David"
<david.lindsey@mindspring.com> wrote:

>Solid suggestions all around. Thanks very much. The first thing I
>have tried is setting the Delay DRAM Read Latch at 0.5ns instead of
>Auto through SoftMenuIII and I haven't had a blue screen in a full day.
> We'll see if that or other BIOS fine tuning does the trick. Fingers
>crossed.
>
>DaveW wrote:
>> When you say that the system is now unstable, I think you've just showed
>> that your "fix" for getting WinMe to use more than 512 MB of RAM is wrong.
>
>I added the "MaxPhysPage=3C000" statement to 386Enh and the
>"MaxFileCache=523264" statement to system.ini. Are you aware of any
>other specific things that would help? The party line from MS seems to
>be that WinME will not support more than 1 Gig. I was under the
>impression that meant ranges between 512 and 1024 Meg could be made
>stable.


You probably don't need a ~ 512MB filecache, try a lower
value, like 262144 to 393216.

Indeed, ME should be able to run up to 1GB.