Re: P4P800se and populating all memory slots?

On Fri, 05 May 2006 15:56:47 GMT, nospam@xxxxxxxxxx (Paul) wrote:

Okay, I'll try that when I got some time. I did try setting the
memory manually, but it always kicked me back to 266mhz 2-2-2.

In article <9lnm529mlmjvcptpe1mrjdjhbffhjn62m6@xxxxxxx>, Drake
<takeahike@xxxxxxxxxxxxx> wrote:

All ram is PC3200.

I've tried using 2x512 basic Crucial modules with 2x512 Samsung
modules on the P4P800se board, and it wouldn't go past the bios menu
(I'd change settings, reboot and it would tell me overclocking failed
and take me back to the bios screen).

I then tried the 2x512 Samsung with 2x256 Mushkin Special 2-2-2, and
it would also take me to the bios menu, but I'd then have to reboot
twice to get it to boot XP. And when it did work, it ran the memory
at 266mhz.

Can you run four modules on this board? Doesn't seem like it.

Speaking in general terms, the chipsets of that era do support
four sticks with no problems. I believe it is quite likely the
chipset always uses 2T command rate, based on the fact that the
memory bandwidth doesn't change on my machine, whether two sticks
or four sticks are used.

If they were my sticks, I'd start by looking at the SPD contents.
The SPD is a tiny eight pin chip on the DIMM, and it contains
information the BIOS uses to program the memory controller.
If you use manual memory settings, you can override some of the
parameters in the SPD.

CPUZ has a dump file, and in it is the raw data from the SPD. The
SPD chip sits on a serial bus used for system management, and that
is how the SPD can be read out by the processor. An example of
a dumped table is here:

http://groups.google.ca/group/alt.comp.periphs.mainboard.asus/msg/009e437f0aa09951?dmode=source

Dump Module #1
0 1 2 3 4 5 6 7 8 9 A B C D E F
00 80 08 07 0D 0A 02 40 00 04 50 60 00 82 08 00 01
10 0E 04 08 01 02 20 00 00 00 00 00 3C 28 3C 28 40
20 60 60 40 40 00 00 00 00 00 00 00 00 00 00 00 00
30 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 AC
40 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
50 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
60 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
70 FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF

In the example above, the length field is 0x80 hex or
128 bytes. And yet only the first 64 bytes have intelligent
data stored in them. If a BIOS is forced to read some
improperly filled out information in the SPD, that can
cause problems, like failure to POST.

The interpretation of the fields is defined in this document.
This is an archived copy of the original, which seems to have
moved or been upgraded or something.

http://web.archive.org/web/20030417070529/http://www.jedec.org/download/search/4_01_02_04R11A.PDF

If the Samsung modules are generic, it could be the
Samsung modules are the "odd man out". Why not complete
your test combinations, by trying the Crucial and the
Muskkin ?

Also, if you want, run with just two sticks at a time,
and get the dumped table above for each pair of sticks.
Then post them here, for comments. CPUZ is here:

http://www.cpuid.com/download/cpu-z-133.zip

If this seems like too much trouble, try the following.
Go into the BIOS, set the speed to "Manual". Set the
timing parameters to 3-3-3-8. I assume all the modules
are PC3200 capable, and so DDR400 would be a choice for
a setting. If you can manage to boot with four of
them, then use CPUZ to see whether the BIOS actually
applied the manual settings or not.

In terms of parameters, in the 3-3-3-8 I mentioned above,
slide #24 here shows which manual parameter is which. Click
on "24" in the left column, then use the Pause "||" button
to stop the animation, so you can read it.

http://corsairmicro.com/corsair/products/tech/memory_basics/153707/index.html

HTH,
Paul

.