Blue-Screen issues with GA-X58-USB3 and CMX6GX3MA1600C9

[donholio]

I picked up a Intel i7-950, Gigabyte GA-X58-USB3 Motherboard, and a CMX6GX3MA1600C9 kit on Black Friday and I have been having significant issues with stability ever since.

 

I have a 550W CoolerMaster RealPower Pro Power Supply that I have had no issues with powering my Core2Duo system before upgrading the Motherboard, CPU, and RAM. I checked the Extreme Power Supply Calculator which told me I should have a minimum 430W, recommended 480W power supply.

 

I backed up my data and reloaded Windows 7 x64 on the hard drive as soon as I could. I began having blue-screen issues almost immediately. What's strange is that I rarely had issues at load. Sometimes I would get a blue screen when I was performing multiple tasks (copying data from one drive over the network and opening a browser window, etc.). Most of the blue screens seemed to happen when the machine goes into a power-saving mode.

 

I have reloaded the OS at least 2 times since the first reload (3 total) in the past week. I have been browsing forum posts and trying different voltages for the QPI/Vtt and DRAM voltages. The motherboard recognizes the X.M.P. Profile and gives me 1.65V for the DDR Voltage but only 1.2V for the QPI voltage. I raised that in the voltage settings window to 1.36V. There is also a DRAM Voltage setting there that is defaulted to 1.5V. I tried to leave that as default, but have since pushed that to 1.66V per discussions on this board.

 

I have even tried to run the RAM at 1066 at the same voltage setting the CAS latency to 9 and CMD rate to 2 manually. None of this seems to help with the stability. So I did what I have in the past - MEMTest. It comes back clear with no issues even after 6 hours so I don't think it's the memory, but rather a settings issue.

 

When I run OCCT and it brings all the cores to full power, I see 74C-78C heat. Nothing over 80 ever.

 

I am wondering if I need to go higher to combat the memory going into a lower-used state and possibly lowering the memory voltage. I am really struggling here to try and find that right combination, but I don't want to go too high on voltages and ruin the CPU, Memory, or Motherboard.

 

Here's the CPU-Z Validation page in case that helps.

http://valid.canardpc.com/show_oc.php?id=1520678

 

Thanks in advance

[trackrat]

You could try 11-11-11-30 2T just to see if it helps. Some Intel CPUs/mobos don't like higher RAM voltages. 1.66 V should be plenty but the RAM is safe to 1.70 V AFAIK, if you want to test.

[Malfeitor]

Your CPU-Z page says your Vcore is 1.168v. Even though the 950 is a little closer to the center of the wafer with a x23 multi, 1.168 seems a tad low for the MB turbo'ing past the locked x23 to x24. You say you memtest fine but crash it under load. I wouldn't say there is anything wrong with the memory. If you are getting STOP 101, bump the Vcore to 1.2v. If you are getting STOP 124 bump the VTT, not to exceed calculated 1.4v. My 12gb CMT kit runs great at 1.6v as long as the IMC voltages are balanced. Just a thought.

 

Cheers

[Yellowbeard]

Have you loaded the most recent BIOS?

[donholio]

All - Thanks for your help, just to clarify- the system is fairly stable under load most of the time. When it is, I can play TF2, BF:BC2 and other games without issues. I see problems mostly just working on the desktop- Copying files and launching a web browser, etc.

 

Trackrat - I will try that, but that seems like I am REALLY having to tweak this memory down to support this Mobo / CPU. I am not even overclocking this thing! What's crazy is this thing isn't even stable at 1066. It's also really random.

 

Malfeitor- Before this, I had the bios set to Auto for every voltage I wasn't setting specifically (DRAM to 1.66 and VTT to 1.36). I went in and set the vCore to 1.2 and CPU-Z showed 1.2V at the low load. As soon as I kicked off OCCT the vCore dropped to the lower value you saw in the CPU-Z post. Just guessing, but since the CPU heat is going to 80 on this crap Intel fan, I wonder if it's undervolting the processor to keep the heat in check. I am picking up a new Zalman fan today to fix that issue. This is the first time I have seen a stock fan not fully support the processor at stock speeds. Intel should be ashamed of themselves.

 

The last stop error I got was 0x0000004e - PFN_LIST_CORRUPT when I booted this morning (voltages set and 1600 memory clock from a cold start). When I looked this up, it was related to corrupt memory.

 

Yellowbeard- That was the first thing I did. There is a beta bios there, but it doesn't look like that does anything but assist with SSD compatibility and I run physical platters.

[Yellowbeard]

Yellowbeard- That was the first thing I did. There is a beta bios there, but it doesn't look like that does anything but assist with SSD compatibility and I run physical platters.

 

You many want to give it a try. The public release notes for a BIOS are never going to be complete and it's very possible that the update could include enhancements for memory compatibility.

[donholio]

Yellowbeard - Will do.

 

All- The blue screen errors I have encountered since reinstalling my OS 2 days ago are:

 

0x0000004e (0x0000000000000099, 0x0000000000149059, 0x0000000000000002, 0x0000000000148ed9)

 

0x0000004e (0x0000000000000099, 0x00000000001657fc, 0x0000000000000002, 0x0000000000164bfb)

 

0x0000001a (0x0000000000005003, 0xfffff781c0000000, 0x000000000000b23c, 0xfffff8a00eaa7009)

 

The 0x0000004e (0x0000000000000099 errors are the PFN_LIST_CORRUPT errors. From what I understand, it could be related to the cache, main memory, or disk corruption. I don't have any disk errors in my event log, but I am going to run a chkdsk.

 

The other error 0x0000001a is a Memory Management error that points to corrupt memory. The error was caused by NTFS.sys. I am now wondering if there isn't an issue with a cold solder joint or something on these sticks. Once it warms up, everything is OK.

 

I have attached the mini dump files in case any of you would like to look at them. I looked at them with BlueScreenView to verify the information above. Thanks again.

Dumpfiles.zip

[donholio]

I installed a Zalman cooler on the CPU and temperatures now peak at 54C with that cooler. The Intel stock went up to 82C before I stopped the OCCT test. This did not seem to help my BSOD issues as right after I put the motherboard back into the computer I got a 4e BSOD.

 

I went ahead and swapped out the memory with the vendor today. So far the computer seems to be a lot happier and I haven't had any BSOD issues yet. Even with a cold boot. I will keep an eye on the system. If it's at rest, I see a sub 1V (0.944V) Core voltage on the CPU. It peaks at 1.184V and ends up at 1.152V throughout the test.

 

Thanks for everyone's help on this. I will update this thread later in the week with more info about how things are going.

[Malfeitor]

"If it's at rest, I see a sub 1V (0.944V) Core voltage on the CPU. It peaks at 1.184V and ends up at 1.152V throughout the test."

 

If you can stay STABLE at this voltage, more power to you. Pun intended. Your stop codes do not appear to be core related. I would investigate your MB's IOH/ICH voltages. I still think the Corsair memory is good passing memtest and suspect it is voltage related or as Yellowbeard suggests, BIOS issue related. Corsair memory is never bad ;):

[donholio]

Malfeitor- Yeah. So far so good. When I pushed it to 1.2V in the bios, once it went under load it went back to the ~1.15V level. I am running stock speed on the processor right now.

 

I will look at the IOH / ICH voltages. There have been a lot of changes I wasn't anticipating when I went to DDR3. I have learned a ton in a very short time.

 

I honestly think the memory I got had some kind of issue with the RAM. I have had only a couple issues with ANY Corsair memory I have purchased and They are the only RAM I purchase. The RAM I got was an open-box return that tested OK by the store employees where I got it.

 

The new RAM I exchanged it for is a lot more stable. More like the performance I expect from a Corsair product.

[Malfeitor]

I recently went to an i7 from my trusty 955 Presler that ran at 4.5ghz for the first 2 years 24/7 and 4.27ghz 24/7 for the last 3 years. It still runs today at 4.27 but my 975XBX Vregs are becoming unstable from "age" and the processor is soon to be a conversational keychain (literally) testament of Intel's fine Engineering. I got sort of blind sided by the integrated memory controller built in to the processor. I had to school-up to get my OC stable and it is very important to note the IMC on the i7's has to be set correctly volt/freq wise in order to be stable and happy. I'm not saying the kit is/was not bad, just suprised if it really is. Out of the approximately 64gb of Corsair memory sticks I own or have owned, I only ever had one 1gb DDR2 Pro stick have memtest errors and RAMGUY replaced it in flawless fashion. You will NEVER catch me with anything other that Corsair in my case. The CMT kit I am running benches at insane numbers. No, i don't work for Corsair, - I just think they have a smart business model and products that are stellar.

 

I have since this picture brought my VTT and Vcore to the 1.38v range

 

Cheers

http://i742.photobucket.com/albums/xx65/MalfeitorGai/Misc/Blix_4244.jpg