Memory Errors and BSOD's on Cold Startup

Toggle sidebar Toggle sidebar
M

mgagnier

Junior Member
Jan 5, 2010
13
0
0
Hello All, I recently bought a new Mobo/Cpu/Ram. When I installed Windows 7 everything was fine for a couple days of up time. When I had the computer off for a day, then turning it back on I was receiving BSOD 0x00000024 stop errors in addition to some other stops.

I ran Memtest x84 v4.0 and it was showing bad RAM. I thought I pegged it down to one stick, so I RMA'd that stick. I thought I had solved the problem, as the computer was running fine afterwards (keep in mind that it was now warmed up) The next day, after the computer was off for several hours, I turned it on and when it made it into Windows, Explorer crashed followed System Service Exceptions, and the 0x0000024 stop error also. I ran the Memtest again, and it was showing bad ram again. However, after the computer was on for an hour or so, the computer would boot up normally and Memtest would NOT report any errors.

Today when I got home from work, I tried booting up and I keep getting different BSOD errors:
Memory Management, APC Index Mismatch, IRQL Not Less Or Equal, and so on... Memtest also reports errors. After an hour or so, the computer booted fine and no memory errors occured on Memtest.

I am inclined to think its a Motherboard issue with the computer being cold at first. Any thoughts on this? PSU has been stable with previous mobo/cpu/ram as was HD and Video card. Here are my specs:

Motherboard: Gigabyte P55A-UD4P Bios updated to F6 from F3.
CPU: Intel i860 2.8 ghz
RAM: 4GB Gskill F3-10666CL8D-4GBHK
PSU: Seasonic M12 600w
HD: WDC Black 1TB
Video: EVGA Geforce GTX 275
OS: W7 Pro x64
 
Last edited:
BTRY B 529th FA BN

BTRY B 529th FA BN

Lifer
Nov 26, 2005
15,188
401
126
Your OS install probably got hosed from the bad stick, do a complete reformat & reinstall and see what happens & then retry Memtest 86 v4.0 along with the OS memory diagnostics.
 
Knavish

Knavish

Senior member
May 17, 2002
910
3
81
BTRY B 529th FA BN said:
Your OS install probably got hosed from the bad stick, do a complete reformat & reinstall and see what happens & then retry Memtest 86 v4.0 along with the OS memory diagnostics.
Click to expand...

He said he got a Memtest error *after* replacing the bad stick. The OS may be corrupted, but there is still a hardware error. The PSU is a good guess.
 
M

mgagnier

Junior Member
Jan 5, 2010
13
0
0
Hi again,

To clarify, Memtest fails when the computer is first being used. After an hour, upon a full power off and on, memtest passes. If I turn off the computer for several hours, and try again, the Computer will bsod constantly/mem test will fail.
 
V

VivienM

Senior member
Jun 26, 2001
486
45
91
I'm a friend of the OP - he's on his way to buying a new PSU now.

Just to clarify a little: when he says "cold" startup, it's not cold boot (i.e. from power off), but cold, i.e. low temperature from sitting off for a while. Think of starting a car after a night outside... only the computer equivalent.

My theory: something flakey is conducting power/signals/etc (i.e. to the RAM) better when it's hot. What that something is, who knows? Though we'll find out when he gets back with his new PSU...
 
V

VivienM

Senior member
Jun 26, 2001
486
45
91
I'm with the OP at his house. New PSU, same problem this morning.

We remounted the mobo. Memtest errors on the first boot. We turn it back off, take out one DIMM, memtest fine. Put DIMM back in, memtest still fine. Seems that was enough to warm it up.

So the question is, what part of this rig fails only when cold, but is quite happy once warm?

Time to try a new motherboard?
 
BTRY B 529th FA BN

BTRY B 529th FA BN

Lifer
Nov 26, 2005
15,188
401
126
Ok, you need to get this cold boot-up thing out of your head. Go into your bios and make sure your ram is running at its rated voltage and timings. Mobo bios makers will not go through the tedious work of making pre-configs for every ram kit out there so they'll follow the norms, or a set of guidlines like for the i7 chips and its 1.65v memory max voltage. In my experience with DDR2 and lga socket 775, 50% of the DDR2 kits were not able to run at its 1:1 rated speeds. e.g. 500fsb will equal DDR2 1000Mhz No doubt DDR3 hasn't changed. It's possible you may have 2 bad pairs or sticks...
 
V

VivienM

Senior member
Jun 26, 2001
486
45
91
BTRY B 529th FA BN said:
Ok, you need to get this cold boot-up thing out of your head.
Click to expand...

Hold on a second. My friend the OP built the thing a week ago. It seems to run fine.

The next day, he boots it up, BSODs galore. Memtest failed. He plays with different RAM configurations, it seems to work. He RMAs 2 of 4 RAM sticks. Computer runs fine.

Next morning, BSOD again. Boots into memtest. It fails. Turns the thing off, etc. It runs fine. He turns it off and goes to work.

He gets home. Boots it up. BSOD. Memtest fails. A few minutes later, after a reboot, it's running good again.

He goes and gets a new PSU. Puts it in an hour later. Reinstalls Windows from scratch. Works fine. Turns it off.

Next morning, boots it up directly into memtest. Fails. He turns the thing off until the evening.

Same thing in the evening. After a few minutes, it starts to run right. He turns it off. That was last night.

He doesn't turn it on this morning. I get here at about 8PM tonight, boot the thing up with NOTHING connected to the board except ODD, RAM, keyboard, and video card, and it fails memtest in front of my eyes. I turn it off, play with the RAM in different slots, etc., and it keeps working. I put the RAM back in the exact configuration it was in 10 minutes earlier, and now it's working fine.

We are about 99.999999% sure that if we don't fix this, it'll be failing memtest again when he turns it on tomorrow morning.

Don't you see a pattern? I see a pattern of the thing failing in the first few minutes after it's been powered off for a few hours, then the thing working.

And I need to get this cold-boot up thing out of my head?!?!?
 
BTRY B 529th FA BN

BTRY B 529th FA BN

Lifer
Nov 26, 2005
15,188
401
126
Have you tested 1 stick at a time?

What about your system panel on your motherboard? are all the header/pins in correctly?

I had a problem with only 1 stick being recognized on my Asrock board until I took out all of the headers on the system panel and just left in the PWR pin. The panel and diagram in the manual were VERY vague; may have been a coincidence but none-the-less a valid question.
 
Last edited:
V

VivienM

Senior member
Jun 26, 2001
486
45
91
BTRY B 529th FA BN said:
Have you tested 1 stick at a time?
Click to expand...

Yup, that's what I did today.

I get here. We take the whole thing apart because we were going to change the board. I notice the screws were a little weird, so I figure we might as well fix the mounting and try again before opening up the new mobo.

I boot it up the first time into memtest. It fails.

I take out one piece of RAM. Boot it up. It passes.
I take out that piece of RAM, put the first one back in, it passes.
I put both of them back in, in the exact same way they were in before, and it passes again. It's been memtesting since I wrote my post an hour or two ago, just fine.

You might think that that means it's fixed. Problem is, so did my friend the OP a few days ago, until he booted it up the next morning and it was bad.

BTRY B 529th FA BN said:
What about your system panel on your motherboard? are all the header/pins in correctly?
Click to expand...

The only thing in there now is the power switch (and CPU fan). And the 24-pin and 8-pin power. I did the absolute bare minimum after remounting the board.

Before tonight, he had all the things plugged in.

We could try unwrapping the new board, but... if that doesn't work, the OP's lost some money (he can resell a shrink wrapped board for more...), and what do we try next? Different RAM? Different CPU?
 
BTRY B 529th FA BN

BTRY B 529th FA BN

Lifer
Nov 26, 2005
15,188
401
126
I'd say get a new RAM kit. Try it with the old board first before you ruin your return option on the new mobo.
 
Last edited:
V

VivienM

Senior member
Jun 26, 2001
486
45
91
Well, we should have listened to you. But that board was too tempting...

Same problem.

I have to say though, I am impressed with that EVGA P55 board. First board I've ever seen with a padded ATX backplatey thingy...

Now, the OP has been googling. Seems there are people on the g.skill forum reporting the exact same issue (RAM errors when the system is cold) with other RAM from g.skill. So... time for him to get new RAM.
 
n7

n7

Elite Member
Jan 4, 2004
21,281
4
81
Honestly, Memtest is not perfect.

I've seen similar weird stuff like that, & it generally does turn out to be bad RAM...
 
V

VivienM

Senior member
Jun 26, 2001
486
45
91
OP went and picked up 8 new gigs of RAM. Tomorrow morning will be the true test. :)
 
V

VivienM

Senior member
Jun 26, 2001
486
45
91
spinejam said:
sounds like my crucial ballistix experience (DDR2)
Click to expand...

I have a rule (that the OP didn't follow in this case... despite hearing it from me for years :)): never buy any kind of RAM with heatspreaders.

Of course, I'm not an overclocker. I just want something I can put in, not have to fiddle with BIOS settings, that'll run properly, and not give me any trouble.

OCZ screwed me once with their overvolted heatspreader junk, lesson learned. Back to Kingston ValueRAM and whatever crucial/kingmax's value brands are...

Funny thing is, up here in the Great White North, it's INCREASINGLY difficult to find good solid JEDEC-spec-following RAM. 95%+ of what retailers carry is overvolted heatspreader stuff...
 
M

mgagnier

Junior Member
Jan 5, 2010
13
0
0
Hi Everyone,

Thanks for your posts and assistance throughout this very long and annoying process.

The culprit was the RAM, in not one of the 2x2 packages of the GSkill DDR3-1333 CL8, but both packages. The problem was just occuring on cold boot. It cost me a bit of money in the end, but I am glad that my system does not have any crap Gskill ram anymore.

FYI - there are lots of postings on the Gskill forums regarding the "cold boot" problem. (See: http://www.gskill.us/forum/showthread.php?t=2479&highlight=cold+boot) Gskill blames motherboards and power supplies, however the issue occurs only on 1156 motherboards with this Gskill stuff. Very difficult to diagnose this problem, since you could only see it happen once every 8 hours or so because the computer has to get nice and cold and be off for that length in order for it to happen again. Word of warning when buying Gskill ram paired with 1156 motherboards.

Glad to be up and running.
 
Z

zagood

Diamond Member
Mar 28, 2005
4,102
0
71
Wow...random. Thanks for biting the bullet on this one and keeping us updated.
 
C

cyclo

Junior Member
May 5, 2007
15
0
0
I had a similar problem with 4 x 2 GB GSkill Ripjaws RAM (8 GB Total) on a Gigabyte P55A-UD4p motherboard. Random BSODs and warning messages from the BIOS when booting up which ultimately resulted in a corrupted the Win 7 OS install. I got the GSkills rated at 1333 (PC10666) 8-8-8-24 timings. I am not into overclocking so I figured I will just set the rated timing in BIOS (SPD sets the timings to 9-9-9-24 after selecting Optimized Settings in BIOS)... I left the voltage at 1.5V (actual reading was 1.58V).

I ultimately ended up replacing the 8 GB GSkill with 8 GB Kingston Value RAMs (1333, 9-9-9-24) and I had no more issues.
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom