BSOD on Vista x64 with 8GB RAM installed

[timgaunt]

I would start with the default timing on 8GB of RAM. Get it running stable.

Thanks SCSIraidGURU however as you will see from my previous posts, when I run at the default timings (1066) the PC can't even boot into windows, the only way I've managed to get it to boot is by downclocking the RAM

 

[SCSIraidGURU]

I would check to see if you need a BIOS rev upgrade. Check to see if your board has a limitation on the number of sticks vs bus speed. My Tyan boards had an issue with more than 6 sides of RAM, 4 double sticks would require DDR333 instead of DDR400. I would also make sure that RAM is the proper spec to run 8GB. Some boards require a different spec of RAM to go above a certain amount. Some require only single sided RAM.

 

[SCSIraidGURU]

Q9550 is 1333 FSB like my E5430 Xeons. I use DDR667 RAM for my Xeons. Most computers with that CPU use DDR800.

 

[jcgriff2]

Sorry for the silence guys but shortly after I posted it really died and one of the sticks on testing was totally dead so I wanted to send that back and get my new sticks before testing anything.

With all the settings in the BIOS reset to the standard settings (i.e. all Voltages set to Auto) and all sticks installed (4 x 2GB on a board that supports 16GB), when starting up I get a selection of the following error messages just before the login screen of a standard boot:

PFN_LIST_CORRUPT
Technical info: STOP: 0x0000004E (0x0000000000000099, 0x000000000008B978, 0x0000000000000000, 0x0000000000000001)
---------------
mountmgr.sys
PAGE_FAULT_IN_NONPAGED_AREA
Techincal info: STOP: 0x00000050 (0xFFFFFA5FFF9D4048, 0x0000000000000000, 0xFFFFFA60009CD9B6, 0x0000000000000005)
mountmgr.sys - Address FFFFFA60009CD9B6 base at FFFFFA60009C2000, DateStamp 47919871
---------------
When attempting to log into Safe Mode I get past the login screen and then get the following BSOD:
IRQL_NOT_LESS_OR_EQUAL
techincal info: STOP: 0x0000000A (0x0000000000000000, 0x0000000000000002, 0x0000000000000001, 0xFFFFF80002094D25)
---------------

0x4e w/ 1st parm 0x99 = PTE or PFN corrupt

0x50 = page fault - mountmgr.sys prob cause..?
P3 (mountmgr) referenced memory add in P1; P2 = 0x0 = read

0xa = MS or kernel-mode driver improperly accessed paged memory
parm 3 = 0x1 = write; p4 accessed memory in P1 = 0x0
This one interesting b/c safemode


I would like to run mini kernel dump files - can't hurt
copy them out of c:\windows\minidump to documents folder or folder of choice

Run msinfo32 and save as NFO file
START | type msinfo32 - save w/ default NFO file extension

Zip up dumps & NFO file and attach to post.

RAM undoubtedly primary suspect - but did you run chkdsk /r ?
RAM increase = page file increase. Is page file system managed?

Regards. . .

jcgriff2

.

 

[sygnus21]

Guys,

If his memory controllers are being stressed, he's going to have problems, no matter the RAM!!!

He didn't need to do anything but relieve the stress on his memory controllers by downclocking the memory since all his dimm slots were filled. Corsair themselves will tell him the samething.

Once he took the stress off the memory controllers his issues went away!!!

In my last post you will notice that I have taken the suggestions from Jim R and sygnus21 of "downclocking" the RAM to 875 I think it was from the factory default of 1066 and so far the system has been running fine

Sorry but he doesn't need to run checkdisk or any other diagnostic programs.

 

[SCSIraidGURU]

Most users don't want to hear the obvious. I would say he needs to set the timing to default and live with it.

 

[mfrey0118]

Wow..I would love to know how to adjust my RAM voltages and such...I guess you could call me a tech-savvy amateur...no education but I am very good with computers by nature.

I am running Toshiba Satellite A215-S6816 with Turion 64X2 AMD dual core and 4gb of RAM and Vista Ultimate x64 and I kept getting BSOD after coming out of sleep mode for ATI driver and audio.
After almost giving up I finally tried taking out the 2GB RAM card I bought (and the store swore to me it was compatible...it's a Kingston) and re-inserting the 1GB that came with the computer and that was it.

But if I can adjust RAM voltage and such, maybe I can make it work...if not how do I know which brands are compatible with each other as far as RAM? Toshiba put two different cards in with the 2GB and 1GB and they work together...help?

 

[SCSIraidGURU]

I have never adjusted a voltage for RAM. Its a overclocking trick for CPUs. You want to match RAM for best performance and reliability. Mixing RAM is not a good idea on most computers. Is your board dual channel? If yes, you want matching sticks to get dual channel to work. My Supermicro board is quad channel, you needed 4 identical sticks per channel. I install 8 x 2 GB identical sticks for best performance.

 

[jcgriff2]

I agree w/ all. Diag probably useless - but after seeing several BSODs elsewhere w/ 6 & 8GB systems, found 2 that turned page file off - didn't believe necessary.

It would just be interesting to see dumps. Driver listing may provide information helpful in preventing future BSODs not related to subject at hand.

Best wishes to all.

 

[Killacore]

Wow, thanks for your speedy replies guys, it's much apprechiated.

I'll re-run memtest tonight as that'll take a while, in regards the error messages, I've had a fair few, isn't there a way I can look through the dump reports for the exact numbers? I did take a couple of photos before but at the time I was running without a PageSys file (see Toms Hardware) so I've just checked but there wasn't any error message though the stop code was 0x0000001E if that helps?

I can recall seeing a less or not_equal a fair few times, othertimes it has been page_fault_in_a_nonpage_area.

These are some shots from my original memtest:

This one just kept looping the erros and didn't go anywhere after 1min 55s (I left it overnight)

Again, left overnight, this was the one that I sent back



Tim

I think there is a strong chance it's your motherboard, If you noticed at the top left hand corner of memtest, it cannot detect your L2 Cache and it incorrectly detects how fast your memory is "MB/s" I have the same processor and an older version of memtest and memtest can detect my memory speed properly and detect my L2 cache.

 

[sygnus21]

I have never adjusted a voltage for RAM. Its a overclocking trick for CPUs

Actually it's not.

Most performance RAM is set to run at higher voltages, around 2.0 - 2.2 range. Most motherboards will only auto adjust to 1.8 - 1.9, and also auto default the timings.

Example: My old Corsair DHX 2gig Twinpack was rated at 2.1volts, 4-4-4-12 timings, yet my Intel D975XBX MB set the RAM at 1.8volts, 5-5-5-18 timings. I had to manually set the correct volts and timings, even though I wasn't overclocking. I also had to do this with my Corsair Dominator Twinpack 4gig kit. It voltage is 2.1v.

With this scenaro it's very easy to run into problems if the memory is not getting enough voltage.

Corsair can confirm this for you.

With the i7 systems, this may change, seeing how the RAM voltage supposedly set at 1.65 volts.

 

[timgaunt]

I think there is a strong chance it's your motherboard, If you noticed at the top left hand corner of memtest, it cannot detect your L2 Cache and it incorrectly detects how fast your memory is "MB/s" I have the same processor and an older version of memtest and memtest can detect my memory speed properly and detect my L2 cache.

hmm, you're not the first person to have suggested that, any ideas on how I can test that? Don't really want to have to rebuild my machine (and have it offline for x days) without being semi-sure.

ATM it's running stable and has been for a few days, temp wise it seems to stay below 50C now so I think adjusting the NB voltages may be done the trick but I've been working short days so haven't had it under any real strain yet...

Tim

 

[SCSIraidGURU]

I would verify the memtest software fully recognizes your chipset and CPU. I have seen it do strange things like this on motherboards not fully compatible to it.

 

[ltwally]

I'll throw in my two cents worth. I want to say that I'm not trying to offend you or your system before-hand, though.

Lesser motherboards often do not have solid enough voltage regulators and/or resistors to handle all memory slots being populated. Your symptoms will be pretty exactly like if you had defective RAM. However, if your motherboard is at fault and simply cannot handle that many sticks of RAM, no amount of replacing your sticks will fix anything.

Heck - even if you bought a high-end motherboard from a quality manufacturer, that doesn't mean that a resistor didn't pop its top or that a regulator isn't failing.

As long as your memory meets the spec of the motherboard, and you're not over-clocking anything.... If you've replaced your memory and have updated your BIOS, I'd seriously start to suspect your motherboard.

Just food for thought.