Asus Sabertooth-X79 Ram issue

[Cyborg-HAF]

I have built a new system using the Asus Sabertooth X79
motherboard and got a 16Gig of Corsair Dominator GT DDR3 2133 ram to
use in the sytem. When got the system ready to power up and explored
my Bios I found in the Tool menu that my D1 slot was Abnormal instead
of OK like the other 3 sticks. When I installed my Windows 7 64-Bit,
I found in System Properties that I only had 12 Gig of total Physical
memory installed instead of the 15 Gig of my set. As a test I swapped
sticks D1 and C1 and posted it again into the Bios and the same D1
slot was still Abnormal so the ram was good and the D1 slot was bad on
the motherboad. Is it possible to populate the ram insead into the
A2, B2, C2, and D2 to avoid the bad D1 slot, the ram would be
recognised by the Bios and all show up. The motherboard manual shows
those slots are used for going to all 8 slots populated instead of
just using 4 sticks. Would I damage the board further by mounting the
ram in number 2 slots? I hate to have to RMA the motherboad over one
bad slot; I have Corsair HD80 closed watercooling set up and it would
be a major hassle to rip everythiing including my processor out and
kill my sysem until get a replacement motherboard.

[Paul]

If a memory configuration is not supported, the BIOS will
beep an error code.

Electrically, I don't see a reason for damage to
occur. The quad channel architecture, is like four
separate memory buses, with two DIMMs per bus. Like this.

--- D1 --- D2

I can't find any rationale for the manual to be filling A1-B1-C1-D1
versus A2-B2-C2-D2.

There is a chipset feature, used on server boards or business laptops,
which lets the IT department reset crashed computers remotely. That
uses a microcontroller in the chipset. And that microcontroller
"steals" a small amount of memory from system memory, in order
to run its firmware load. In such a case, the microcontroller
does not have the same flexibility as the main CPU, and may require
a particular DIMM slot to be filled first. I cannot find any
reference to this in the documentation for this system (X79/LGA2011).

http://en.wikipedia.org/wiki/Intel_Active_Management_Technology (AMT)
http://en.wikipedia.org/wiki/VPro

Paul

[Cyborg-HAF]

I moved my four Dimms to the A2, B2,C2, and D2 channels, but
just got Beep codes and no post. I had to move the ram back to the #1
channels to get the system to post, and back to the D1 channel
Abnormal and loss of 4 Gig of ram. I have to remove my radiator each
time I move ram, so it was a real pain. I guess need to RMA the
motherboard, which will be nightmare ripping all my cooling hardware
and processor, ram, fan cables, power cables, etc. I guess I have to
RMA throungh NewEgg rather than try to do this through Asus.
Herb

[Paul]

Thanks for the feedback on the #2 RAM slots.

I can see a few reports, with Asus LGA2011 motherboards, with a particular
slot not working. But at this point, haven't seen a thread where the
situation improved.

One person had working RAM the first time, and on subsequent days
had some of the RAM disappear. So whatever connection problem
existed for that person, it didn't happen the first time.

It's possible there is a BIOS component to this problem.

There was a similar problem with LGA1366 triple channel RAM.

In terms of the land grid array processors, when they've been
inserted once, you can pull the processor and examine the gold
pads on the processor. You should see a "dot" in the center of
each gold pad, caused by the land grid "spring" biting into
the gold. If you examine the bottom of the processor, and you
see a quadrant of the processor with very light or non-existent
bite marks, that would imply a problem with the socket.

There have been differences in the past, in the quality of socket
designs. Some Foxconn brand sockets had problems with making
contact. The Lopes brand sockets were supposed to be better.
But that was some time ago, and with one of the lower pin count
sockets. I haven't seen any web site articles about socket
problems with LGA2011 generation. A couple of users are
finding a single bent pin in their socket area, but they weren't
sure whether it was there when the board was delivered or not.

So the searches I've done so far, haven't identified a "smoking gun"
for missing RAM. I don't know what to suggest next for you,
except to suggest it isn't actually a RAM problem, but something
else. The RAM sticks are probably fine, and it's something else
about the DIMM slots themselves, the LGA2011 socket versus CPU,
or some kind of BIOS issue. You could try another version of
BIOS - that is, if the system runs well enough to take a chance
on flashing the BIOS.

Paul

[Paul]

Well, one user here got relief, by changing memory brands. Initial problem
was only two of four DIMMs were recognized.

http://vip.asus.com/forum/view.aspx?SLanguage=en-us&id=20120326225527758&board_id=1&model=SABERTOOTH%20X79&page=2&count=12

And in terms of recognizing RAM, there is a difference between BIOS
recognition, and what is shown in CPUZ. In CPUZ, the program queries
the SPD chip on the DIMM in each slot. This is a relatively
easy test to pass, because all that is happening is slow speed
serial communications with the SPD chip.

The BIOS on the other hand, for it to "recognize" a DIMM, not only
does it read the SPD (to get memory size, speed, and so on). But,
it also uses the ancient "probing" technique to verify the RAM
size. If the "probing" tecnnique is failing (no location
exhibits reliable storage), the BIOS rejects the DIMM and
it won't show up in the memory map. So even though we live
in the "SPD era" and the tables in the SPD tell the BIOS
what to do, the BIOS does a double check. And the result
of this careful approach, is the BIOS generally does not
approve RAM which is less than 100% operational.

If you use CPUZ, it will show all the DIMMs. And all that's needed
for that to happen, is for the SPD pins to make contact, and for
the SPD chip to have power applied. Lots of data pins might not
be making contact, and CPUZ could still "read out" the DIMM. But the
BIOS needs the vast majority of the pins to be working, to give
a good RAM indication.

But as I said in the other post, the odds are this is not a RAM
or RAM bus issue. It's some other kind of issue. And it may not
be resolved by reseating the CPU.

In the example thread above, the posts by "Pookieraaj", it could
either be that the RAM is defective or substandard, or, that the
BIOS did not do a good job of handling the SPD info, and DIMMs
were rejected for the wrong reasons. A new BIOS can sometimes
sort that out.

Paul