RAM modules from different vendors within the same PC?

[Giovanni Azua]

Hello all,

I have been complaining to the DELL technical support
because I found I was installed in my new Workstation
four 512MB RAM modules from different vendors:

I have 2GB = 4x512M

2x512MB from Hyundai
2x512MB from Samsung

Does anyone know for certain if a system could become
error prone i.e. potential Windows blue-screens or
whatever side-effect like loss of optimal performance
because of having disparate RAM modules installed within
the same system? Is it possible to find out any effect
of this?

I thought about doing the following "test case":

1-. Leave all four modules, run PassMark PerformanceTest
and collect results.

2-. Remove the two Samsung RAM modules, run PassMark PerformanceTest
and collect results.

3-. Put back the two Samsung RAM modules and remove the
two Hyundai, run PassMark PerformanceTest and collect results.

4-. Finally compare performance results for all the three test fixtures.

I would like to argue better my complain to DELL.

Thanks in advance,
Best Regards,
Giovanni

 

[Mike Walsh]

Different brands of memory should not conflict with each other. Any memory that is compatible with your BIOS configuration and motherboard chipset should work equally well. There have been reports of cases is this NG and others when various configurations e.g. more than 2 DIMMs will not work properly. Some memory will cause problems because timing is not compatible with motherboard settings, or because the memory is defective.

 

[Apollo]

Hello all,

I have been complaining to the DELL technical support
because I found I was installed in my new Workstation
four 512MB RAM modules from different vendors:

I have 2GB = 4x512M

2x512MB from Hyundai
2x512MB from Samsung

Does anyone know for certain if a system could become
error prone i.e. potential Windows blue-screens or
whatever side-effect like loss of optimal performance
because of having disparate RAM modules installed within
the same system? Is it possible to find out any effect
of this?

I thought about doing the following "test case":

1-. Leave all four modules, run PassMark PerformanceTest
and collect results.

2-. Remove the two Samsung RAM modules, run PassMark
PerformanceTest
and collect results.

3-. Put back the two Samsung RAM modules and remove the
two Hyundai, run PassMark PerformanceTest and collect
results.

4-. Finally compare performance results for all the three test
fixtures.

I would like to argue better my complain to DELL.

All will probably be fine, performance is not really the issue,
stability could be and is very easy to test. Any variance in
performance between the two makes will most likely be extremely
small.

I don't really know what PassMark is, I assume it's some kind of
stress test? Memtest86 is what you really want, it quite often
shows memory errors where other tests say all is fine.

Run memtest86 for a couple of hours minimum, if you get any errors
reported then try your options 2 & 3, again running for a couple
of hours each. Write down all the errors you get (if any) as
you'll want to pass this information onto the Dell techs and ask
them what they propose to do about it.

As I said, all will most likely be fine, but the first thing I do
when taking receipt of a new PC is run memtest, I've seen a few
with dodgy memory from new.
http://www.memtest86.com/ download the ISO image and plonk it onto
a cd, boot from the cd and your off.

 

[kony]

Hello all,

I have been complaining to the DELL technical support
because I found I was installed in my new Workstation
four 512MB RAM modules from different vendors:

I have 2GB = 4x512M

2x512MB from Hyundai
2x512MB from Samsung

Does anyone know for certain if a system could become
error prone i.e. potential Windows blue-screens or
whatever side-effect like loss of optimal performance
because of having disparate RAM modules installed within
the same system? Is it possible to find out any effect
of this?

I thought about doing the following "test case":

1-. Leave all four modules, run PassMark PerformanceTest
and collect results.

2-. Remove the two Samsung RAM modules, run PassMark PerformanceTest
and collect results.

3-. Put back the two Samsung RAM modules and remove the
two Hyundai, run PassMark PerformanceTest and collect results.

4-. Finally compare performance results for all the three test fixtures.

I would like to argue better my complain to DELL.

Thanks in advance,
Best Regards,
Giovanni

You weren't necessarily entitled to all-identical memory and
have no complain so long as the system works properly.

Your tests are not going to reveal anything useful unless
machine was VERY instable. Try running a memory test
instead, memtest86+ (Google will find it).

It is common, perhaps even to be expected, that the memory
benchmark scores will go down with all 4 modules, but not
necessarily because they're different, rather that there's 4
instead of 2, 3.

What I would complain about instead is that they'd used 4
memory slots to end up with 2GB, rather than a 2 x 1GB
module configuration which would leave the additional 2
slots available still. That would be MUCH more upsetting to
me on a professional level workstation unless you're 100%
certain you might never need more memory.

 

[w_tom]

Memory tests are not very informative if the memory is
intermittent AND you don't put the memory in an environmental
extreme. For example, the best extreme to locate intermittent
memory is heat. Execute that memory test with the memory
heated by a hair drier on high. Even at those temperatures
(uncomfortable to touch but does not leave skin), the memory
is well within its normal limits. If any intermittent is
going to be found by a memory test, it will most likely be
found when the memory is heated.

BTW numbers provides with system crashes would have
suggested which memory is defective. What did the Dell
comprehensive hardware diagnostics report?

 

[Giovanni Azua]

Hi,

What I would complain about instead is that they'd used 4
memory slots to end up with 2GB, rather than a 2 x 1GB
module configuration which would leave the additional 2
slots available still. That would be MUCH more upsetting to
me on a professional level workstation unless you're 100%
certain you might never need more memory.

I requested to have the 4x512MB actually, was part of
a very hard negotiation to get a big discount ... actually
DELL at least Switzerland did not offer the 4x512 when I
bought this one, just 2x1GB which would have costed
me additional 1000USD

Best Regards,
Giovanni

 

[Giovanni Azua]

Hi Ian,

I downloaded and left memtest86 running the whole night 8.7 hours
and there was not a single error found. I think I will have to start
accepting the fact of having different RAM brands :-( but well
if it works DELL is right, I should basically not care.

Thank you for your help after I had memtest86 checking I feel
better about it.

Best Regards,
Giovanni