Memtest86 settings?

In message: 1c1ip0t2d7haco6mogak2df85g39g5n4qj.RemoveThis@4ax.com,
Mike <dy73ui39aHe5.RemoveThis@hotmail.com> typed:

 > I always use the default settings. I only use memtest86 in the
 > event I've been experiencing blue screens or lockups. My system is
 > normally very stable indeed, 99% of the time past problems have
 > nearly always been the cause of bad memory. It last occurred when I
 > added a second 512meg stick of twinmos memory, memtest showed there
 > was errors in some memory locations, I swapped it out for a
 > replacement stick and it then passed all the tests okay, and no more
 > blue screens. )

Well that's just it. My system works fine with any memory setting, but this
damn Memtest just bugs me... I mean, logic dictates that a good memory stick
should pass any test without throwing out errors no matter how agressive
memory testing is. Or does it? I really need some more feedback on this.
Say, would you try and test your memory with all Memtest settings set to max
(like "test all memory") and see if it starts acting up. You don't even have
to run the whole test, it should start throwing out errors almost instantly.

--
"Ad eundum quo nemo ante iit..."<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

 > Well that's just it. My system works fine with any memory setting, but
this
 > damn Memtest just bugs me... I mean, logic dictates that a good memory
stick
 > should pass any test without throwing out errors no matter how agressive
 > memory testing is. Or does it?

I've had a stick go bad without memtest being able to find a problem. I
could run memtest on them for a week straight, and never see an error - but
*any* machine I put the stick in would start having blue screens/seg
faults/lockups/etc.. The vendor (Crucial) replaced the stick with no
questions asked, and everything was fine after that.

 > I really need some more feedback on this.
 > Say, would you try and test your memory with all Memtest settings set to
max
 > (like "test all memory") and see if it starts acting up. You don't even
have
 > to run the whole test, it should start throwing out errors almost
instantly.

Memory failure isn't always as simple as a good or failed cell. Whether
a particular memory cell fails can depend on things like the pattern of data
stored in cells around it. When memtest *does* find errors on bad sticks
for me, it is very rare for the errors to pop up right away - it usually
takes a little while. I wouldn't consider it reasonably sure unless memtest
can run for at least a couple of days without throwing errors.

steve<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

Steve Wolfe wrote:
  >> Well that's just it. My system works fine with any memory setting,
  >> but this damn Memtest just bugs me... I mean, logic dictates that a
  >> good memory stick should pass any test without throwing out errors
  >> no matter how agressive memory testing is. Or does it?
 >
 > I've had a stick go bad without memtest being able to find a
 > problem. I could run memtest on them for a week straight, and never
 > see an error - but *any* machine I put the stick in would start
 > having blue screens/seg faults/lockups/etc.. The vendor (Crucial)
 > replaced the stick with no questions asked, and everything was fine
 > after that.

That's sorta the converse ... a "good" memory stick should pass every test
you can dream up, including memtest86. A "bad" stick may get a pass from
memtest86, but fail in other tests. So if a particular test fails (such as
memtest86) then I would say that the stick is faulty.

Incidentally, there's also power issues in how well memory works. Memtest86
applies little or no stess to the CPU, GPU, hard disks, etc, and the load is
very constant. This is the "best case" scenario for memory voltage supply.
Under "normal" operating conditions, such as running Prime95 or 3D games or
similar, the power draw from the rest of the system is heavier, and you get
more noise on the supply lines from the constantly changing load. This can
push borderline memory (that is OK when nothing else is loaded) to the point
of failure.

[...]

--
Michael Brown
<a style='text-decoration: underline;' href="http://www.emboss.co.nz" target="_blank">www.emboss.co.nz</a> : OOS/RSI software and more
Add michael@ to emboss.co.nz - My inbox is always open<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

In message: YNemd.2680$9A.99849@news.xtra.co.nz,
Michael Brown <see.TakeThisOut@signature.below> typed:

 > That's sorta the converse ... a "good" memory stick should pass every
 > test you can dream up, including memtest86. A "bad" stick may get a
 > pass from memtest86, but fail in other tests. So if a particular test
 > fails (such as memtest86) then I would say that the stick is faulty.

So... I guess i should return it then right? But I worry, what if the new
stick behaves in the same manor? Now at least I have two *identical* sticks,
who knows what i'll end up with if i opt for a new one... Damn it. Should I
return it? Again, I must stress that I had *no* blue screens or shit like
that, teh memory seems to work fine, but there's always that little voice
that says it will mess up my data in some way sooner or later. And I had
just such problem with a faulty memory stick in the past, which really
messed up some of my files while copying, introducing massive errors. This
data was later irretrivable...

--
"Ad eundum quo nemo ante iit..."<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

 >In message: YNemd.2680$9A.99849@news.xtra.co.nz,
 >Michael Brown <see DeleteThis @signature.below> typed:
 >
  >> That's sorta the converse ... a "good" memory stick should pass every
  >> test you can dream up, including memtest86. A "bad" stick may get a
  >> pass from memtest86, but fail in other tests. So if a particular test
  >> fails (such as memtest86) then I would say that the stick is faulty.
 >
 >So... I guess i should return it then right? But I worry, what if the new
 >stick behaves in the same manor? Now at least I have two *identical* sticks,
 >who knows what i'll end up with if i opt for a new one... Damn it. Should I
 >return it? Again, I must stress that I had *no* blue screens or shit like
 >that, teh memory seems to work fine, but there's always that little voice
 >that says it will mess up my data in some way sooner or later. And I had
 >just such problem with a faulty memory stick in the past, which really
 >messed up some of my files while copying, introducing massive errors. This
 >data was later irretrivable...

Before returning it, have you tried increasing the memory voltage a extra notch
or 2, my Twinmos memory gives loads of errors at the default 2.5v buts its fine
when its upped to 2.6v or more.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

In message: c6tjp0p8iq3h8qthqodrsl9sdrqkfvi9u7.DeleteThis@4ax.com,
Mike <dy73ui39aHe5.DeleteThis@hotmail.com> typed:

 > Before returning it, have you tried increasing the memory voltage a
 > extra notch or 2, my Twinmos memory gives loads of errors at the
 > default 2.5v buts its fine when its upped to 2.6v or more.

Thnx, I will try that...

--
"Ad eundum quo nemo ante iit..."<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

 > Before returning it, have you tried increasing the memory voltage a extra
notch
 > or 2, my Twinmos memory gives loads of errors at the default 2.5v buts its
fine
 > when its upped to 2.6v or more.

If a stick wouldn't live up to its rated values at the rated voltage, I'd
return it in a heartbeat - even if an extra .1 volt did mask the problem.

steve<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings? 

On Tue, 16 Nov 2004 16:34:16 +0100, "Dr.Gonzo" <drgonzo.TakeThisOut@hi.htnet.hr>
wrote:

 >In message: c6tjp0p8iq3h8qthqodrsl9sdrqkfvi9u7.TakeThisOut@4ax.com,
 >Mike <dy73ui39aHe5.TakeThisOut@hotmail.com> typed:
 >
  >> Before returning it, have you tried increasing the memory voltage a
  >> extra notch or 2, my Twinmos memory gives loads of errors at the
  >> default 2.5v buts its fine when its upped to 2.6v or more.
 >
 >Thnx, I will try that...
Also try another memory tester.

There is a "Gold memory test" around that is good though I have not tried
it beyond the BX Chipset.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Memtest86 settings?