SOLVED Strange RAM problem on the M4A88T-M motherboard

[edtate]

Hi guys,

Spec:

M4A88T-M Motherboard
AMD 960T Black Edition
Corsair Hydro series H80
OCZ Stealth Xstream II 600W PSU
HD 6870 GPU
Kingston 8GB (2x4GB) DDR3 1600Mhzs HyperX CL9 1.65

My computer seems to run fine 24x7 until I start to play games. What happens in that after a period of intensive usage is that it eventually restarts with an 0x124 error (hardware fault) after about an hour.

Now the weird thing is that when I swapped RAM slots from blue to black (Asus recommends putting the RAM in the blue slots for overclocking purposes), the system runs with more stability, in that it takes quite a while longer to get to the 0x124 BSOD.

Initially I thought the problem was due to the graphics card since the error occurs whilst playing games, trying all possible driver combinations did not lead to the error though, and the BSOD does not mention any system files relating to the ATI driver. Also, the error occurs whilst doing a context switch of some kind, (like alt-tabbing mid-game or clicking a UI button) which is why I thought it may be a RAM issue.

The RAM is set to 1600 in the BIOS, and the clock timings are those recommended by ASUS, 9-9-9-9-27 with 1.65v voltage (I bumped the voltage from 1.5v, however at 1.5v I was still getting this crash).

So I clocked my RAM down from 1600 to 1333 and set all to auto and still got the error.

Running memtest86 on the RAM overnight revealed no apparent errors.

I also set the NB frequency to 2400mhz (3xDRAM freq) but this didn't stabilise anything.

Now I'm kind of at a loss as to what the error could be, but I suspect that the memory controller in the CPU may be causing the fault (since apparently 980t supports only 1333mhz RAM ??).

Is there some way of verifying / stabilising the CPU memory controller, or does anyone have any better ideas?

Thanks in advance

 

[redduc900]

With 8GB of RAM installed running at DDR3-1600 and 9-9-9-27 1.65V, you may need to bump the CPU/NB voltage approx. +0.10V to gain stability.

 

[trents]

Give us some screen shots of CPU-z tabs: "CPU", "Memory" and "SPD". This may or may not be a ram problem. It could be unstable because of an undervolted CPU, for instance. I will also tell you that that recent AMD CPUs do not like that high volt ram you are running. The ram that will do 1600 mhz at 1.5 volts works much better.

 

[edtate]

Give us some screen shots of CPU-z tabs: "CPU", "Memory" and "SPD". This may or may not be a ram problem. It could be unstable because of an undervolted CPU, for instance. I will also tell you that that recent AMD CPUs do not like that high volt ram you are running. The ram that will do 1600 mhz at 1.5 volts works much better.

The CPU hasn't been overclocked yet, I wanted to get the computer stable before doing any overclocking. Voltage is at stock 1.264~ set to auto in the BIOS. I'm pretty sure the problem has nothing to do with the voltage of the CPU in this case since turning off cool 'n quiet and running CPU bound code which bumps the utilised voltage doesn't cause a crash.

Setting CPU-NB to +0.10V (from 1.15V to 1.25V) seems to have fixed the problem and I can no longer reproduce the crash! Thanks alot for your help reducc you have solved my problem!

 

[edtate]

I spoke too soon. The CPU-NB voltage increase of +0.10V caused some instability which led to be putting it at 1.2V. Still getting the crashes now though.

 

[edtate]

Additional notes on the problem itself:

The screen blacks out (music still plays in bg) and then the computer eventually resets itself after about an hour of intensive gaming. There is no BSOD, but the event viewer displays the Kernel-Power critical error (41 I think). At first I thought it was a BSOD but after running Windows under "do not reset after crash" it still blacks out.

So tried a bunch of things with various different results...

Tried increasing CPU voltage to 1.3V. No change.
Tried VDDNB voltage at 1.2 (+0.1V), and 1.25 (+0.15V). No change.
Tried putting RAM in black slots rather than blue slots, the motherboard manual says to put it in blue slots for OC reasons. I put in black and the crash occured later than usual, but still eventually occured.

Replaced RAM with 1333mhz G-Skill RAM to ensure that the 1600mhz RAM wasn't causing the problems. RAM is 9-9-9-24 1T 1333mhz 1.5V. Tried bumping RAM voltage to 1.55V, no change.

Replaced GFX card with one which used only 1 PCIE power connector from the PSU and this seemed to fix the problem.

The weird thing is that I can run the furmark stress test for a while and no crash occurs.

So this leads me to believe that the problem is either with the power supply or the graphics card.

What do you guys think I should do?

 

[edtate]

Arrf. So I replaced the PSU with one from a different machine (a corsair 750TX) and the problem persists.

Black screen after 1hour of gaming... Only happens on PowerColor Radeon HD6870 and not on an older card.

Thinking of RMAing it...

 

[itsmechg]

[SOLVED] How? can't see any next page.