RAM probs under Win 7 64 bit.

[Yousuf Khan]

I've had bad RAM that Memtest86+ didn't detect.

So, what's your point?

That's exactly my point, haven't you been paying attention?

Yousuf Khan

 

[Nil]

That's exactly my point, haven't you been paying attention?

Your claim was that Microsoft's memory tester was in ineffective
"placebo" because Memtest86+ found an error that Microsoft missed. That
doesn't prove anything. All memory testers are liable to miss errors
and give different results - including Memtest86+.

 

[Brian Gregory [UK]]

Your claim was that Microsoft's memory tester was in ineffective
"placebo" because Memtest86+ found an error that Microsoft missed. That
doesn't prove anything. All memory testers are liable to miss errors
and give different results - including Memtest86+.

Memtest86+ is a very intensive test with many separate sections, including
an optional bit fade section that lasts, if I remember correctly, 90
minutes.

I highly doubt that something Microsoft threw together for Windows 7 matches
it.

 

[Nil]

Memtest86+ is a very intensive test with many separate sections,
including an optional bit fade section that lasts, if I remember
correctly, 90 minutes.

I highly doubt that something Microsoft threw together for Windows
7 matches it.

I don't claim that it does, but that doesn't mean that Memtest86+ is
infallible. Nor does it mean that Microsoft's memory tester is "a
placebo."

 

[Gene E. Bloch]

alt.windows7.general:

I don't claim that it does, but that doesn't mean that Memtest86+ is
infallible. Nor does it mean that Microsoft's memory tester is "a
placebo."

BTW, I also have had bad RAM that memtest86+ thought was good...just to
corroborate your remark a few posts up.

 

[Brian Gregory [UK]]

BTW, I also have had bad RAM that memtest86+ thought was good...just to
corroborate your remark a few posts up.

I have seemingly had that too, but only on particular type of Dell PC at
work.

I decided (possibly wrongly I admit) that the way the chipset was configured
was wrong so that Windows was somehow changing the RAM timings slightly,
perhaps when it put the video display in graphics mode, (those PCs share
system RAM as video RAM).

 

[Paul]

I have seemingly had that too, but only on particular type of Dell PC at
work.

I decided (possibly wrongly I admit) that the way the chipset was configured
was wrong so that Windows was somehow changing the RAM timings slightly,
perhaps when it put the video display in graphics mode, (those PCs share
system RAM as video RAM).

Windows doesn't change RAM timing. And the memory setup is done by the
BIOS, such as setting Top Of RAM register. That is why it is very
important for the BIOS to know how to detect and configure RAM. And
also detect and configure bus address space for the add-in cards and chips.

It's also why the BIOS uses two methods for memory detection. Not only
does it use the SPD declaration ROM on each DIMM, but it double
checks that declaration, by doing the traditional peek/poke method
as well. One user had an incorrectly programmed SPD chip (wrong DIMM size
programmed in it), and the computer still ran fine. And that was
because the BIOS detected the situation correctly and only set the
decoder to point to the RAM that was really there.

There are utilities that allow reprogramming the memory parameters
while running Windows. It's just the OS itself, that doesn't
delve into such things.

Windows has things like MAXMEM, but that is an after-the-fact
restriction on the memory available within Windows. It isn't a
reprogramming of the hardware.

One other disparity you can have, is the numbers shown for CAS and
the like, in the BIOS screen, can be different than what the BIOS
has entered into the controller. Using CPUZ, you can verify the
values used. I had one motherboard, where it might show "CAS = 3"
in the BIOS, and then when in Windows, you'd see "CAS = 2". And this
was the BIOS doing something different than the interface showed. By
doing a BIOS update, that behavior stopped. So what you see, isn't always
what you get. And is why you have to be a bit careful, to verify
your hardware with something like CPUZ, to see if the BIOS isn't
doing its job properly. If you're building computers for a living,
checking the BIOS settings match CPUZ, should be part of your process.

Paul

 

[GreyCloud]

Well the move from 1066 to 1333MHz he thought had fixed the prob did not
last more than 3 or 4 restarts.

He's now done the Memtest86 on all 2Gb modules and found one giving numerous
errors.....oh dear....back to the supplier.

And don't forget to purchase an anti-static kit.
Otherwise, you may damage the new ram stick.

 

[Brian Gregory [UK]]

Windows doesn't change RAM timing. And the memory setup is done by the
BIOS, such as setting Top Of RAM register. That is why it is very
important for the BIOS to know how to detect and configure RAM. And
also detect and configure bus address space for the add-in cards and
chips.

It's also why the BIOS uses two methods for memory detection. Not only
does it use the SPD declaration ROM on each DIMM, but it double
checks that declaration, by doing the traditional peek/poke method
as well. One user had an incorrectly programmed SPD chip (wrong DIMM size
programmed in it), and the computer still ran fine. And that was
because the BIOS detected the situation correctly and only set the
decoder to point to the RAM that was really there.

There are utilities that allow reprogramming the memory parameters
while running Windows. It's just the OS itself, that doesn't
delve into such things.

Windows has things like MAXMEM, but that is an after-the-fact
restriction on the memory available within Windows. It isn't a
reprogramming of the hardware.

One other disparity you can have, is the numbers shown for CAS and
the like, in the BIOS screen, can be different than what the BIOS
has entered into the controller. Using CPUZ, you can verify the
values used. I had one motherboard, where it might show "CAS = 3"
in the BIOS, and then when in Windows, you'd see "CAS = 2". And this
was the BIOS doing something different than the interface showed. By
doing a BIOS update, that behavior stopped. So what you see, isn't always
what you get. And is why you have to be a bit careful, to verify
your hardware with something like CPUZ, to see if the BIOS isn't
doing its job properly. If you're building computers for a living,
checking the BIOS settings match CPUZ, should be part of your process.

Paul

My suspicion was that the BIOS had messed up the hardware register values of
the highly dodgy VIA chipset and that when Windows changed to the graphics
mode the RAM timings changed a little because the chipset was either buggy
or incorrectly programmed.

 

[Yousuf Khan]

Your claim was that Microsoft's memory tester was in ineffective
"placebo" because Memtest86+ found an error that Microsoft missed. That
doesn't prove anything. All memory testers are liable to miss errors
and give different results - including Memtest86+.

It's completely your own decision to hide your head in the sand, if you
like.

Yousuf Khan

 

[Yousuf Khan]

BTW, I also have had bad RAM that memtest86+ thought was good...just to
corroborate your remark a few posts up.

Then the next question would be, did that RAM passed by memtest get
detected by Microsoft's tester instead?

Yousuf Khan

 

[Yousuf Khan]

My suspicion was that the BIOS had messed up the hardware register values of
the highly dodgy VIA chipset and that when Windows changed to the graphics
mode the RAM timings changed a little because the chipset was either buggy
or incorrectly programmed.

Those old VIA chipsets were a lot of fun, nothing ever seemed to work
the way they were meant to. Every new feature introduced seemed to come
with side-effects. It sounds like we're talking about a disease.

Yousuf Khan

 

[Gene E. Bloch]

Then the next question would be, did that RAM passed by memtest get detected
by Microsoft's tester instead?

Yousuf Khan

It didn't exist at the time.

I may have used another memory tester too (DrMem or something, I
forget), but the way I solved the problem was to remove sticks until I
found the bad one, since the symptoms were *very* indicative of bad
RAM, in spite of what the RAM tester(s) said.

 

[Gene E. Bloch]

It's completely your own decision to hide your head in the sand, if you like.

Yousuf Khan

I'm afraid that in this case Nil's answer reflects reality a lot better
than yours.

 

[Nil]

It's completely your own decision to hide your head in the sand,
if you like.

My decision is to ignore unsubstantiated claims by internet blowhards.

 

[Jumbo Jack]

And don't forget to purchase an anti-static kit.
Otherwise, you may damage the new ram stick.

He/we were using an wrist earth band thanks

All fixed now, faulty module exchanged and system running 100%

 

[KCB]

He/we were using an wrist earth band thanks

All fixed now, faulty module exchanged and system running 100%

Jumbo Jack,
It might be worth your time to manually delete the previous poster's
signature in your replies. It is the two dashes with a following space, and
whatever is below that. Some news readers automatically trim this, which,
in effect, would also delete your reply, making it look like you
posted...nothing.