yes, I did post this to the AMD forum but that forum looks dead (three
weeks since last post) so I post this here. I think this is an easy
question to answer...

I recently got an old PC that has a AMD Sempron 140 chip. I upgraded
the memory to DDR3 (just more of the same as what was in the original)
but now the system is unstable. I don't think it's a virus, so it's
the memory settings.

What is a conservative BIOS setting for this chip to deal with DDR3
memory, I think the memory is 1066 MHz Unganged Mode (2x 2GB sticks)
but I have a feeling I need to half that to 535 MHz. I have an
American MegaTrends Bios from around 2009/2010.

Sorry for being so vague; I can get more info if anybody responds to
this. Just some conservative memory settings would be appreciated.
it's set to "Auto" now but the system is unstable (reboots at odd
times).

PS--I did run MemTest86 for a long time and it reports no errors in
the new memory... but not sure if MemTest86 runs memory slower than
the bios or what.

PPS--now here's the stuff from the American Megatrends BIOS screen:

Advanced Chipset Setup:

Dram Frequency: AUTO (maybe set it to "Limit"--I am guessing one
problem is that DDR3 has a 'default' 1024 MHz speed while this old AMD
Sempron 140 chip wants 535.8 MHz or half the speed? Just a guess)

Dram Timing: AUTO
DCT Unganged Mode: AUTO
Share Memory Auto Detection: AUTO

Now there's a chance my problem is not memory but a virus, since this
HD has a bunch of rogue programs on it, but the instability only
started after I installed more memory---so I don't think so.

Any ideas? Like I say I did run MemTest86 for a whole day and it
found no errors in the new memory.

OS is Windows 7. Memory is 4 GB.

RL

On Mar 25, 2:41*am, RayLopez99 <raylope...@gmail.com> wrote:

>
> Any ideas? *Like I say I did run MemTest86 for a whole day and it
> found no errors in the new memory.
>
> OS is Windows 7. *Memory is 4 GB.
>
> RL

It's kind of funny that if I "unlock" the chip in BIOS it is
misidentified on startup as a AMD Athlon II X2 4400e, instead of a
Sempron 140. So I went back to "locking" it.

RL

On Mar 25, 2:41*am, RayLopez99 <raylope...@gmail.com> wrote:
>

I got a reply from the AMD board, and I replied there:

Are the two sticks identical in brand and specification?
Post screenshots of CPU-Z.
Compare what CPU-Z says for your memory settings when using each of
the two stick alone, do they differ?

Test also with the new stick alone, does it still appear being
stable?
Try also with setting the RAM frequency manually to 800MHz DDR3, in
BIOS.




Thanks man! That was quick. I had originally calendared to check this
thread in a week's time. The reason I'm back: to report that what I
did was simply reseat the sticks, and that seemed to "solve" the
problem, for now.

But you may be onto something with the two sticks. You see, I bought
from a seedy merchant what seemed to be the same sticks (he tells me)
but I notice the serial numbers and even the form factors are off (one
is a "half height" stick that has the second top hole kind of cut
off).

So I'll check CPU-Z for each stick and repost here.

On Mar 25, 3:32*am, RayLopez99 <raylope...@gmail.com> wrote:
> On Mar 25, 2:41*am, RayLopez99 <raylope...@gmail.com> wrote:
>
>
>
> I got a reply from the AMD board, and I replied there:
>
> Are the two sticks identical in brand and specification?
> Post screenshots of CPU-Z.
> Compare what CPU-Z says for your memory settings when using each of
> the two stick alone, do they differ?
>
> Test also with the new stick alone, does it still appear being
> stable?
> Try also with setting the RAM frequency manually to 800MHz DDR3, in
> BIOS.
>
> Thanks man! That was quick. *I had originally calendared to check this
> thread in a week's time. *The reason I'm back: *to report that what I
> did was simply reseat the sticks, and that seemed to "solve" the
> problem, for now.
>
> But you may be onto something with the two sticks. *You see, I bought
> from a seedy merchant what seemed to be the same sticks (he tells me)
> but I notice the serial numbers and even the form factors are off (one
> is a "half height" stick that has the second top hole kind of cut
> off).
>
> So I'll check CPU-Z for each stick and repost here.

seems the sticks were not seated properly...and this caused an
occasional fail. Weird. Either that, or I got some virus that's
undetected. But I think it was the improperly seated sticks.

RL

In response to a poster on the AMD board who suggested using CPU-Z to
check each memory stick was identical (which they were):

Here are the screenshots, and as you can see the two sticks are
identical. So it seems that 'reseating' the chips 'solved' the
problem; perhaps they were loosely connected. But it's strange that
MemTest86 did not detect any problem despite running it for nearly one
day.

Well, I cannot seem to attach a screenshot to this post, but you'll
have to believe me that the sticks were identical. The only field
different in CPU-Z was "Channels #" which was "Dual" with both sticks
in, and "Single" with each stick in separately. The DRAM Frequency was
535.8, the FSBRAM was 3:8, the CAS# Latency (CL) was 7.0 clocks; the
RAS# to CAS# Delay (tRCD) was 7 clocks, the RAS# Precharge (tRP) was 7
clocks, the Cycle Time (tRAS) was 20 clocks, the Bank Cycle Time (tRC)
was 27 clocks, and the COmmand Rate (CR) was 1T. This was with the
BIOS set to "AUTO" and Unganged (the default) for the Bios.

Other settings:
Advanced Chipset Setup:
Dram Frequency: AUTO
Dram Timing: AUTO
DCT Unganged Mode: AUTO
Share Memory Auto Detection: AUTO


I guess the lesson of this thread is that an improperly seated memory
stick can pass all memory tests like the excellent Memtest86 test and
yet still fail in practice--if indeed this problem was the result of
an improperly seated memory stick.

RayLopez99 wrote:
> On Mar 25, 2:41 am, RayLopez99 <raylope...@gmail.com> wrote:
>
>> Any ideas? Like I say I did run MemTest86 for a whole day and it
>> found no errors in the new memory.
>>
>> OS is Windows 7. Memory is 4 GB.
>>
>> RL
>
> It's kind of funny that if I "unlock" the chip in BIOS it is
> misidentified on startup as a AMD Athlon II X2 4400e, instead of a
> Sempron 140. So I went back to "locking" it.
>
> RL

If your system is unstable with the second core unlocked, then lock it
again.

http://www.cpu-world.com/CPUs/K10/AM...0HBGQBOX).html

Test the memory, one stick at a time first.

Unganged may be appropriate for a quad core and multiple threads.
On a single core, with the auto setting, perhaps the BIOS will
choose a ganged setting.

In addition to using memtest86+ (memtest.org), you also have
the option of running Prime95 (www.mersenne.org/freesoft). It
is a multithreaded program. With both cores enabled, it would
run two testing threads. With the processor locked again,
there would be one test thread. Prime95 should run for at least
four hours, without any of the test threads stopping. If a test
thread stops, something is still wrong.

Prime95 does not identify bad memory locations. It tells you
"something" is wrong with the system, but not what is wrong.
As such, it is a "system acceptance test", a way of saying
"it is Prime95 stable". Memtest86+ on the other hand, is
a diagnostic, which attempts to give a bad memory address.
But because memtest86+ isn't a stress test as such,
other test cases you can construct, are more stressful.

That is why, both tests are useful.

Paul

RayLopez99 wrote:

>
>
> I guess the lesson of this thread is that an improperly seated memory
> stick can pass all memory tests like the excellent Memtest86 test and
> yet still fail in practice--if indeed this problem was the result of
> an improperly seated memory stick.

Now run Prime95, and get back to me in four hours :-)

You're not done yet. Passing memtest86+, is nothing.

Paul

On Mar 25, 7:08*am, Paul <nos...@needed.com> wrote:
> RayLopez99 wrote:
>
> > I guess the lesson of this thread is that an improperly seated memory
> > stick can pass all memory tests like the excellent Memtest86 test and
> > yet still fail in practice--if indeed this problem was the result of
> > an improperly seated memory stick.
>
> Now run Prime95, and get back to me in four hours :-)
>
> You're not done yet. Passing memtest86+, is nothing.
>
> * * Paul

Thanks Paul I learned that this chip is in fact a dual core chip with
one core disabled, hence the "locked" attribute switch in BIOS.

Also I did download the prime number generator program you mentioned,
ran it for an hour, but since my memory problem is already solved, I
gave up when it showed no errors. I did not realize that an
improperly seated memory stick can pass memory tests and yet fail in
the real world. That was a surprise to me.

RL