PC3200 Memory Wierdness

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On Tue, 01 Feb 2005 20:42:43 GMT, root@andyc.carenet.org (root) wrote:

>I had something a little wierd happen the other day that I cannot explain
>and noone that I have talked to can either. I put together a new computer,
>which used DDR400/PC3200 memory. One of the chips was bad, so I used
>the chips from my old computer, which was DDR266. When I got the chip
>replaced, I put them in, and on a whim ran memtest (this is a Suse Linux
>system). Well, in the span of 2 or 3 minutes, I had 14 errors. I then
>realized that I had not changed it to DDR400. Thinking it would be even
>worse, I changed it to DDR400, and no errors. I let it run for 3 hours,
>and no errors. It was always my understanding that you could back the
>speed off on memory, and that was all well and good. Any ideas why
>something like this would happen?

There *are* limits to how far you can back off but I'd agree that I
wouldn't expect running DDR400 memory at DDR266 to be too much, though
someone more expert than I may chime in with a correction.:-) My guess is
that with the DDR400 rating, some weirdness between SPD and BIOS
interpretation of it, caused the timings to be set too low for the memory
to work properly... e.g. if the DDR400 CasL was 3 and the BIOS decided that
for DDR266, CasL = 2 should work just fine.

--
Rgds, George Macdonald

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

In article <34j1015g1quh2c4hf7146quod7r31d4ml1@4ax.com>,
George Macdonald <fammacd=!SPAM^nothanks@tellurian.com> writes:
> On Tue, 01 Feb 2005 20:42:43 GMT, root@andyc.carenet.org (root) wrote:
>
> There *are* limits to how far you can back off but I'd agree that I
> wouldn't expect running DDR400 memory at DDR266 to be too much, though
> someone more expert than I may chime in with a correction.:-) My guess is
> that with the DDR400 rating, some weirdness between SPD and BIOS
> interpretation of it, caused the timings to be set too low for the memory
> to work properly... e.g. if the DDR400 CasL was 3 and the BIOS decided that
> for DDR266, CasL = 2 should work just fine.
>

That makes sense. I should be able to figure that out if I try it
again. Thanks!

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

On Tue, 01 Feb 2005 20:42:43 GMT, root@andyc.carenet.org (root) wrote:

>I had something a little wierd happen the other day that I cannot explain
>and noone that I have talked to can either. I put together a new computer,
>which used DDR400/PC3200 memory. One of the chips was bad, so I used
>the chips from my old computer, which was DDR266. When I got the chip
>replaced, I put them in, and on a whim ran memtest (this is a Suse Linux
>system). Well, in the span of 2 or 3 minutes, I had 14 errors. I then
>realized that I had not changed it to DDR400. Thinking it would be even
>worse, I changed it to DDR400, and no errors. I let it run for 3 hours,
>and no errors. It was always my understanding that you could back the
>speed off on memory, and that was all well and good. Any ideas why
>something like this would happen?

Memory chips can have different sets of timings for different speeds.
It is possibly (almost certain in fact) that the timings used for
DDR266 speeds are faster (in terms of clocks) than those used for
DDR400. It may even be that one or more of the settings is faster in
absolute terms at DDR266 speeds than for DDR400 speeds. As such, it's
quite possible for a stick of memory to run fine at DDR400 speeds but
not at DDR266 speeds. The most likely reason for this is the fact
that many memory module manufacturers do a piss-poor job of
assembling, testing and programming their SPD settings for their
modules.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca