A defected memory paired with working memory installed on mobo does
not indicted memory
error when it boots up. This tricks the PC guru thinking both memory
is working. I've had to argue with a man sold old defected memory
when I returned it. He did not understand why the defected memory
was booted up fine when it paired with working memory. He even
maintained that I've had to install as a pair. I did not know how
to show him the memory he sold was defect it. The system normally
boots up by ignore the defected memory and read only working
memory. He showed me the first boot up screen to prove the memory
he sold me was working and did not read the total memory size in BIOS
program.

Red Cloud wrote:
>
> A defected memory paired with working memory installed on mobo does
> not indicted memory
> error when it boots up. This tricks the PC guru thinking both memory
> is working. I've had to argue with a man sold old defected memory
> when I returned it. He did not understand why the defected memory
> was booted up fine when it paired with working memory. He even
> maintained that I've had to install as a pair. I did not know how
> to show him the memory he sold was defect it. The system normally
> boots up by ignore the defected memory and read only working
> memory. He showed me the first boot up screen to prove the memory
> he sold me was working and did not read the total memory size in BIOS
> program.

He's not a guru but an idiot or a liar because he should know that the
boot-up memory test is not very thorough at all, and most computers
are configured so they perform only an incomplete version of it that
mostly just counts the amount of memory installed in the system.

To really test memory fairly well you need to run at least one of
these diagnostics, each for several hours, and preferrably test with
more than one kind of computer (one with an AMD CPU, another with an
Intel CPU):

MemTest86 www.MemTest86.com
MemTest86+ www.MemTest.org
Gold Memory www.GoldMemory.cz

The first two programs are very similar but have given me different
results. I've seen diagnostics run for 4-5 hours before they reported
errors, and one diagnostic didn't find the same error again for almost
10 hours.

Most memory modules are made with substandard chips that failed some
testing done by the $1,000,000 (US) testing machine used by actual
chip manufacturers. Those chips are not marked with that
manufacturer's part numbers or logo or may be covered with useless and
purely decorative heatsinks. Most memory module makers perform their
own, much more lax tests on those chips, often using nothing but
motherboards. To always avoid getting bad memory, buy modules made
only with chips marked as having come from Samsung (SEC), ProMOS,
PowerChip, Nanya, Inotera, Elpida, Micron, or Nynix. As far as I
know, the only ways to be guaranteed of this is by either looking at
the chips on the modules, or by purchasing Samsung or no-heatsink
Micron/Crucial modules.

On Jan 1, 12:42*am, "larry moe 'n curly" <larrymoencu...@my-deja.com>
wrote:
> Red Cloud wrote:
>
> > A defected memory paired with working memory installed on mobo does
> > not indicted memory
> > error when it boots up. *This tricks the PC guru *thinking both memory
> > is working. * I've had to argue *with a *man sold old defected memory
> > when I returned it. He did not understand *why the defected *memory
> > was booted up fine when it paired with working memory. He even
> > maintained *that I've *had to install as a pair. * I did not knowhow
> > to show him the memory he sold *was defect it. *The system normally
> > boots up by ignore the defected memory and read only working
> > memory. * *He showed me the first boot up screen to prove the memory
> > he sold me was working and did not read the total memory size in BIOS
> > program.
>
> He's not a guru but an idiot or a liar because he should know that the
> boot-up memory test is not very thorough at all, and most computers
> are configured so they perform only an incomplete version of it that
> mostly just counts the amount of memory installed in the system.
>
> To really test memory fairly well you need to run at least one of
> these diagnostics, each for several hours, and preferrably test with
> more than one kind of computer (one with an AMD CPU, another with an
> Intel CPU):
>
> * * *MemTest86 * * * * *www.MemTest86.com
> * * *MemTest86+ * * * *www.MemTest.org
> * * *Gold Memory * * * *www.GoldMemory.cz
>
> The first two programs are very similar but have given me different
> results. *I've seen diagnostics run for 4-5 hours before they reported
> errors, and one diagnostic didn't find the same error again for almost
> 10 hours.
>
> Most memory modules are made with substandard chips that failed some
> testing done by the $1,000,000 (US) testing machine used by actual
> chip manufacturers. *Those chips are not marked with that
> manufacturer's part numbers or logo or may be covered with useless and
> purely decorative heatsinks. *Most memory module makers perform their
> own, much more lax tests on those chips, often using nothing but
> motherboards. *To always avoid getting bad memory, buy modules made
> only with chips marked as having come from Samsung (SEC), ProMOS,
> PowerChip, Nanya, Inotera, Elpida, Micron, or Nynix. *As far as I
> know, the only ways to be guaranteed of this is by either looking at
> the chips on the modules, or by purchasing Samsung or no-heatsink
> Micron/Crucial modules.

Never use memtest86 and no need to use it. Setup program in BIOS
displays size of the working memory.

Red Cloud wrote:
> On Jan 1, 12:42 am, "larry moe 'n curly" <larrymoencu...@my-deja.com>
> wrote:
>> Red Cloud wrote:
>>
>>> A defected memory paired with working memory installed on mobo does
>>> not indicted memory
>>> error when it boots up. This tricks the PC guru thinking both memory
>>> is working. I've had to argue with a man sold old defected memory
>>> when I returned it. He did not understand why the defected memory
>>> was booted up fine when it paired with working memory. He even
>>> maintained that I've had to install as a pair. I did not know how
>>> to show him the memory he sold was defect it. The system normally
>>> boots up by ignore the defected memory and read only working
>>> memory. He showed me the first boot up screen to prove the memory
>>> he sold me was working and did not read the total memory size in BIOS
>>> program.
>> He's not a guru but an idiot or a liar because he should know that the
>> boot-up memory test is not very thorough at all, and most computers
>> are configured so they perform only an incomplete version of it that
>> mostly just counts the amount of memory installed in the system.
>>
>> To really test memory fairly well you need to run at least one of
>> these diagnostics, each for several hours, and preferrably test with
>> more than one kind of computer (one with an AMD CPU, another with an
>> Intel CPU):
>>
>> MemTest86 www.MemTest86.com
>> MemTest86+ www.MemTest.org
>> Gold Memory www.GoldMemory.cz
>>
>> The first two programs are very similar but have given me different
>> results. I've seen diagnostics run for 4-5 hours before they reported
>> errors, and one diagnostic didn't find the same error again for almost
>> 10 hours.
>>
>> Most memory modules are made with substandard chips that failed some
>> testing done by the $1,000,000 (US) testing machine used by actual
>> chip manufacturers. Those chips are not marked with that
>> manufacturer's part numbers or logo or may be covered with useless and
>> purely decorative heatsinks. Most memory module makers perform their
>> own, much more lax tests on those chips, often using nothing but
>> motherboards. To always avoid getting bad memory, buy modules made
>> only with chips marked as having come from Samsung (SEC), ProMOS,
>> PowerChip, Nanya, Inotera, Elpida, Micron, or Nynix. As far as I
>> know, the only ways to be guaranteed of this is by either looking at
>> the chips on the modules, or by purchasing Samsung or no-heatsink
>> Micron/Crucial modules.
>
> Never use memtest86 and no need to use it. Setup program in BIOS
> displays size of the working memory.

The BIOS memory test doesn't work right.

I had an Nforce2 chipset motherboard, with dual channel memory,
and the BIOS memory test could not detect an entire dead chip!
That's how bad the BIOS test is. It's not guaranteed to catch anything.

Memtest86+ boots from a floppy, CD, or USB stick. If tests everything
except BIOS reserved areas (perhaps 1MB worth). As such, it covers about
as much as you can reasonably expect to cover.

Memory test programs which run while the OS is running, are limited in
test coverage, to the free RAM area.

Paul

In message
<d08c2ed4-80c7-45e0-8fab-(E-Mail Removed)>
someone claiming to be Red Cloud <(E-Mail Removed)> typed:

>Never use memtest86 and no need to use it. Setup program in BIOS
>displays size of the working memory.

Displaying the size of the memory available to the BIOS and testing to
see if it works are very different things.

--
It's always darkest before dawn. So if you're going to
steal your neighbor's newspaper, that's the time to do it.

On Jan 1, 11:38*pm, Paul <nos...@needed.com> wrote:
> The BIOS memory test doesn't work right.
>
> I had an Nforce2 chipset motherboard, with dual channel memory,
> and the BIOS memory test could not detect an entire dead chip!
> That's how bad the BIOS test is. It's not guaranteed to catch anything.
>
> Memtest86+ boots from a floppy, CD, or USB stick. If tests everything
> except BIOS reserved areas (perhaps 1MB worth). As such, it covers about
> as much as you can reasonably expect to cover.
>
> Memory test programs which run while the OS is running, are limited in
> test coverage, to the free RAM area.
>
> * * Paul

Yes, I recently installed Ubuntu Linux on a machine that I'm sure I
have tested thoroughly in the past and thought was fine. Memtest86+
was included in the Ubuntu GRUB menu and it found an error that occurs
only at two particular pseudo random values.

On Jan 2, 4:39 am, DevilsPGD <Still-Just-A-Rat-In-A-C...@crazyhat.net>
wrote:
>
> Displaying the size of the memory available to the BIOS and testing to
> see if it works are very different things.

I get it, now. If the bad memory shows tested properly, then buy a
new BIOS chip and replace it!

In message
<6779064c-cce3-4d2f-b2e0-(E-Mail Removed)>
someone claiming to be Flasherly <(E-Mail Removed)> typed:

>On Jan 2, 4:39 am, DevilsPGD <Still-Just-A-Rat-In-A-C...@crazyhat.net>
>wrote:
>>
>> Displaying the size of the memory available to the BIOS and testing to
>> see if it works are very different things.
>
>I get it, now. If the bad memory shows tested properly, then buy a
>new BIOS chip and replace it!

Exactly!

--
It's always darkest before dawn. So if you're going to
steal your neighbor's newspaper, that's the time to do it.