Sign in with
Sign up | Sign in
Your question

Bad Memory Controller or DIMM Slot

Last response: in Motherboards
Share
August 28, 2010 9:33:56 PM

Hello,

The last few weeks, my PC has been struggling to boot and randomly power off for the last few weeks. It may suddenly power off after typing in the password when the PC does boot into windows. After that, the performance was sluggish until Win 7 loads up everything. Other times, it would randomly blue screen.

I did benchmark the CPU, RAM, and the video card and they all passed the stress test (Cinebench, POV Raytrace, Fritz chess, Prime95, Linx, Everest, Furmark, wPrime etc). Also, I tried two different video cards and PSU's.

Today, it refuses to power up.

I tested the RAM by installing into each slot to see if it booted up and post. After testing each slot and each of the 4 sticks, I determined that DIMM slots 1 and 3 (dual channel) causes random reboots or no post. So, I am running 4gb on DIMM 2 and 4 and that runs fine with no problems.

Does this mean I have a bad memory controller on the CPU? Or, the DIMM slots are no longer good?




Intel i7-875k (1156) CPU
Gigabyte P55A-UD7 Motherboard
G.Skill 8gb (4x 2gb) DDR 3 2300mhz RAM
EVGA 1gb GTX 260
Kingwin 1000 Watt PSU (LZG-1000)
MS Win 7 (64-bit)
a c 716 V Motherboard
August 28, 2010 10:36:16 PM

First guess is MOBO. How may HDD/SSD on your rig? Any in RAID? Also, what, if any, BIOS or software changes to the system? Any OC'ing? If so what are all of the settings?

It "feels" like a controller is overheating and causing the BSOD and/or immediate shutdowns and/or erratic behavior. Software, can easily cause BSOD, but typically won't do the other symptoms ~ few exceptions (e.g. Flash 10.1 or GA software can render similar behavior).
m
0
l
August 28, 2010 11:27:00 PM

3 HDD, 2 ODD, nothing in raid, and no software instability.


The i7-875k is OC to 4ghz (24 X 167 @ 1.38 V). The ram is running at 2004 mhz @ 1.6 V. No other voltages were changed.
m
0
l
Related resources
a c 716 V Motherboard
August 28, 2010 11:35:09 PM

Simple first - disconnect the 2 HDD leaving the {primary}; just unplug @ MOBO. Run any HDD bench, e.g. ATTO, on the single HDD. Q - what controller(s) is/are the HDD attached? If there's still a failure clock down both the CPU & RAM to spec {save current OC config as profile first}.

Also, just to humor me - uninstall all Flash components and if installed any GA Utilities.
m
0
l
August 28, 2010 11:59:47 PM

I've ruled out PSU, video card, RAM, ODD, and the HDD. There's nothing wrong with any of the SATA ports or the USB ports. Only DIMM 1 and 3 are non-functional.

Win 7 Professional has not experienced any problems nor any of the chipset (Nvidia, motherboard, etc).
m
0
l
a c 716 V Motherboard
August 29, 2010 12:10:24 AM

Multi-HDD have a proven problem with GA MOBOs - so I take it that you tried the single HDD? The old Flash left alone the GPU and relied on the CPU, but beginng in 10.1b1 Flash {accesses GPU} has also been proven to do everything that you've stated. So,I am not running you on some frivolous goose chase.

The ports themselves is/are not the concern, it's MOBO power management of those "ports."

Since you've R/O all of the components then we're back to my {"First guess is MOBO"}.

Reconfirm @ http://forums.tweaktown.com/ and ask for Lsdmeasap {he's online}.
m
0
l
a b V Motherboard
August 29, 2010 12:34:21 AM

This sounds to me as if it is a power supply problem, but you said that you have ruled out the power supply as being the cause of the problem. How did you rule out the power supply as being the cause of the problem? The only way of being sure that it is not the power supply is to swap it with a known good one, simply measuring the voltages is not good enough. If you have PROVED that it is not the power supply then the next most likely candidate is unfortunately the motherboard.
m
0
l
a c 716 V Motherboard
August 29, 2010 12:37:01 AM

The only other explanation is bent CPU pins {can cause "missing" memory, but that's not what you described - http://www.intel.com/support/processors/sb/CS-030850.ht...

Above, my concerns are OC and power instability which can compound or cause the the problems. Since I've asked that you (simple) R/O by 1. lower controller power consumption and 2. possible OC instability / power consumption, and you don't want to buy into my reasoning - that's okay. My "Goal" is to find the problem and make sense of it all later.

BTW - I did read 2 or 4 DDR3 are not working ~ to test I did mean for you to install ALL 4 RAM sticks, and I "assumed" all 4 have the same exact OC settings and are the same exact P/N {part numbers} and most importantly are 100% confirmed compatible with your MOBO.
m
0
l
August 29, 2010 1:49:46 AM

Hello,

@ pjmelect:

The PSU is not the problem as I tested two quality PSU's before. FYI, I finally tested the DIMM slots one by one and found DIMM 1 and 3 do not work properly.


@jaquith:

Yes I tried to raise and lower the voltage, but nothing happens. At 4ghz the voltage needs to be 1.38 to 1.4 depending if you are OC by multiplier (like me) or by base clock.

All ram sticks are the same name, manuf, P/N, etc. They all passed memory benchmarks and work fine in the working DIMM slots. They are not part of the Gigabyte approved list as those lists tend to be outdated, but regardless the RAM sticks had no problems for half a year on this board and previous boards (Intel and AMD).
m
0
l
a b V Motherboard
August 29, 2010 5:45:56 AM

Motherboard it is then, bad luck. :(  You may be able to fix the motherboard though, look at the motherboard for bulging tops of capacitors or capacitors not standing up straight due to the bottoms of the capacitor bulging. If you are good with a soldering iron it is easy to replace them. (Replace with 105 C types). This probably won’t fix the bad memory slots but might fix the computer starting up problem.
m
0
l
a c 716 V Motherboard
August 29, 2010 6:15:12 AM

Yeah, I OC plenty and it is a trial by error for most to achieve a fully stable and reliable system. Any OC above 4GHz especially more by error. I have no idea if your settings are okay or not {reliable & solid}, but as you know when you begin to see an instability a light-bulb {should} goes off to examine if the settings are indeed "really" okay which is why I recommended "trying" stock (save your settings a profile and 1 boot later you're right back where you left off).

I've found a higher BCLK with a lower CPU multiplier actually works better, and I try to keep my voltages as low as possible. Most of my builds are on EVGA, and 2 identical rigs can have totally different results. Also, I don't push my DDR3 too much and in some cases not at all if 2-3 FPS cause a rig to crash 2-3 hours into a game or whatever processing you're doing - the gain is insignificant.

Most importantly, I am not trying to preach or be critical but as I said help.

However, the first 4 words of my post stands. BUT "what if" different settings work?

Good Luck!
m
0
l
a b V Motherboard
August 29, 2010 7:23:22 AM

Admiral_Cecil said:
Hello,

The last few weeks, my PC has been struggling to boot and randomly power off for the last few weeks. It may suddenly power off after typing in the password when the PC does boot into windows. After that, the performance was sluggish until Win 7 loads up everything. Other times, it would randomly blue screen.

I did benchmark the CPU, RAM, and the video card and they all passed the stress test (Cinebench, POV Raytrace, Fritz chess, Prime95, Linx, Everest, Furmark, wPrime etc). Also, I tried two different video cards and PSU's.

Today, it refuses to power up.

I tested the RAM by installing into each slot to see if it booted up and post. After testing each slot and each of the 4 sticks, I determined that DIMM slots 1 and 3 (dual channel) causes random reboots or no post. So, I am running 4gb on DIMM 2 and 4 and that runs fine with no problems.

Does this mean I have a bad memory controller on the CPU? Or, the DIMM slots are no longer good?

Intel i7-875k (1156) CPU
Gigabyte P55A-UD7 Motherboard
G.Skill 8gb (4x 2gb) DDR 3 2300mhz RAM
EVGA 1gb GTX 260
Kingwin 1000 Watt PSU (LZG-1000)
MS Win 7 (64-bit)


No, it most likely means you don't know which slots are first, second, third, forth.

Intel 1156 and 1366 boards need to have the 2nd slot fille first, if you're counting over from the CPU. This is unique to 1156 and 1366 boards, and affects all 4-slot 1156 and all 6-slot 1366 boards.

So, the first slot is the second from the CPU. The matching slot for dual-channel is the forth from the CPU. If you had a 1366 socket, the matching slot for triple-channel would be the furthest slot from the CPU.

That is to say, the order to fill slots is:

CPU->Slot 3->Slot 1->Slot 4->Slot 2

For 1156 boards, you install two modules in this order: 0-1-0-1 where 0 is empty and 1 is filled.

You must do that because it's part of the design. Hundreds of complaints on Newegg are from buyers who sent boards back assuming they were defective, because they didn't know that.

This order does not affect AMD or previous-generation Intel platforms, so it's not unique to DDR3. It's only unique to the Core-i processor series.
m
0
l
August 29, 2010 2:09:36 PM

@Crashman:


Hello,

Originally, all slots were filled with memory 4 X 2gb for the last 5 months.

Gigabyte P55A-UD7 goes by 2-1-4-3. I checked the manual for this.

Right now, my memory are in Slots 1 and 3 currently (if I go by the 2-1-4-3). However, if I place memory on 2 or 4 individually -or- all 4 slots, the system reboots randomly or no post at all.


Today I updated to the newer BIOS from F6 to F7 and re-populated all four DIMM slots and everything works w/o any problems.
m
0
l
August 30, 2010 3:44:05 PM

I too wanted to suggest that you power down as many of the controllers as possible until I read what jaquith was over and over recommending you do to isolate the problem by cooling or reducing loads on all the controller temps. If it were me I simply would recommend the you clear CMOS and load optimized settings on a bare bone system.

Gigabyte motherboards are notorious with power and controller overheating issues and just because your over clock was stable or barley stable four months ago doesn't make it so now. Dust, use, curing all play a factor in keeping stability over time.

I would increase your base clock, base clock's of 190-200 seem to work better over time where the target is 4ghz or more.
m
0
l
a c 716 V Motherboard
August 30, 2010 6:21:38 PM

@ i7cpu3d I won't run the i7-875K @ 4 GHz from what I've read; over 3.7 GHz is pushing that CPU {thermal/W spikes @ >3.7}. Sure you can run tests @ 4 GHz+, but that's completely different than 7/24/365; maybe on an EVGA. The OP's rig "seems" unstable as configured, and yep that's what I was trying to get across. I'm guessing the MOBO is more less toast. Also, if you read something and copy it thinking it is Gospel often leads to failure, people who know how to OC often end up with odd BCLK #'s 173 or 194, etc to "find" out the alchemy of settings, but my intention is not to criticize the OP instead to share info.
m
0
l
!