BSoD on 3 month old build

[jpsenior]

Hey everyone.

I've been having some trouble with a new build and i kind of abandoned the post for the past month because I've been away. I built a new rig about 3 months ago and it's been having some problems. Crashing frequently with the same BSoD: IRQL Not Less or Equal. here's a picture:

[/img]

To summarize, i've run memtest for about 90 minutes, i'm not sure how many "passes" it was because it was my first time using the program and i was unfamiliar with it, however, it found no errors. I have also run SFC /scannow. No problems were found with the integrity of any of the files.

What is my next step?

Thanks in advance (sorry if i shouldn't have made another forum post about this)!

 

[mdsiu]

System spec's please

 

[jpsenior]

of course. my bad.

1. AMD A8-3850
2. Gigabyte GA-A75-UD4H FM1 AMD A75
3. OCZ ModXStream Pro 700W
4. G.SKILL Ripjaws Series 4GB (2 x 2GB) 240-Pin DDR3
5. Radeon HD 6850 1GB 256-bit GDDR5 PCI Express 2.1 x16
6. Western Digital Caviar Blue WD5000AAKS 500GB

 

[gamerk316]

IRQL_NOT_LESS_OR_EQUAL_TO is almost always due to bad RAM, though a faulty device driver could be at fault.

What I find VERY odd is the memory address referenced: 0x00000008, which should be an address reserved for use for hardware addressing. The fact an address that low is causing a problem strikes me as odd. So it could be a driver or hardware issue rather then RAM. [That being said, I once had a stick of RAM where EVERY even bit failed...]

To clarify, aside from the HDD's and GPU, there is no other hardware installed in the PC?

 

[jpsenior]

all 6 specs in my previous post comprise the entire computer. 1 hard drive, 2 sticks of ram, gpu, psu, cpu, mobo, dvd drive.

 

[pat]

Try to give the RAM a little bit more voltage. Gigabyte motherboard seems to underpower RAM, at least, in the one that I tried. Once the RAM get more voltage, complete stability was achieved. At least, in my case.

 

[jpsenior]

how do i go about doing that? and what are some possible ramifications of doing so?

 

[13thmonkey]

all 6 specs in my previous post comprise the entire computer. 1 hard drive, 2 sticks of ram, gpu, psu, cpu, mobo, dvd drive.

Over and above that there will be built in hardware that all use memory locations, soundcard, network card, usb, serial (if its there) etc. So Gamerk316, it could be a bad mobo?

 

[pat]

how do i go about doing that? and what are some possible ramifications of doing so?

It is done in BIOS. Look for something like DDR voltage or RAM voltage. Open the MIT option, then system voltage control. You'll see DDR3 voltage. Now, your memory is set to run at 1.5 volts. I would try 1.55V at first, then going higher but no need for higher than 1.65V. If it cannot be stable then the problem is elsewhere.

By giving a little bit more voltage, this help in case the PSU or voltage regulator drop a little under specs. This can then cause data corruption and crash.

 

[pjmelect]

You did not say which version of memtest86 you were running to test your memory. I have found that version 4.0 is very much better at detecting faulty memory than version 3.5 and that if you are using version 3.5 I am not surprised that it found nothing wrong.

 

[pjmelect]

how do i go about doing that? and what are some possible ramifications of doing so?

I liked the pun

 

[jpsenior]

haha thanks.. but yeah i ran memtest version 4.0a so, no problems there?

and is upping the voltage for my RAM a safe option? how will i know if it stabilizes the system? just if it doesn't crash? because that's my only problem right now. frequent crash+bsod.

 

[PurpleHayes]

Upping the voltage on RAM will, like with anything else, increase the heat output by the RAM, but as long as you stay below 1.65V you'll be perfectly safe (the increase in heat won't be very significant).

To test the system stability, running MemTest and watching for BSODs is all I can think of at the moment (I'll look for some benchmarking software or something), but definitely do what pjmelect says and up the voltage a bit; if your BSODs go away, problem solved. If not, you'll need to start considering hardware problems

 

[pat]

Yes, it is sate to increase the voltage as there is some memory that run 1.65v, some goes as high as 1.8v. If you don't get anymore crash with the voltage increase, then you should have fixed your problem.

Memtest is good to test memory for errors, but since it has to run from DOS, there is no load in the system as when you are playing a game. The 3d chip, as well as the HDD and sound each ask for power, and it can cause the power to fluctuate. especially the 3d chip.

 

[jpsenior]

i appreciate all the help. i'll try it out later tonight

 

[gamerk316]

I wish I could remember what addresses that low are mapped to; its been ages since I've had to deal with that stuff. But I'm 99% sure that the address thats giving the fault (0x00000008) is reserved for hardware addressing...

Do you have another GPU to test with? Also, might want to do a chkdsk, just to make sure its not a HDD problem [i doubt it though].

 

[jpsenior]

no i don't have another gpu. The first one i got came dead on arrival, this is the second one. It's been fine except for when i first installed it i got a blue screen that told me the drivers were not installed correctly. but then i uninstalled and reinstalled the drivers and never had that issue again.

I'm gonna try increasing the voltage on the ram momentarily, i'll report back later to tell you guys if it crashed again.

 

[jpsenior]

I just upped the voltage to 1.550. I never tried to configure speeds or anything like that. is that something i should look into.

Furthermore.. Are there other hardware tests i should be running?

 

[pat]

Do as you do to have crashes happen and see. if it is stable, then your system was underpowering the RAM, or the RAM was receiving less power than it needed.

 

[jpsenior]

got it. i did it yesterday afternoon and no crashes yet, but it's barely been a full day. so i'll keep up with this thread with more updates later. I appreciate everyone's help thus far.

 

[pat]

When I first set up my current Gigabyte motherboard, and knowing that I have 4 x 4GB modules, I put the memory voltage up when I setup the BIOS for the first time before instaling the OS. Never bothered to look back if l could set it as default as it is rock stable as it is.

 

[jpsenior]

So the computer crashed again last night. The RAM is still at 1.550v. Should i up it to 1.600v? or run some diagnostic beforehand? or a new strategy?

 

[gamerk316]

New strategy I think...

I can't rule out bad drivers, or RAM at this point. Not having another GPU to test with also complicates things slightly.

Just for kicks, uninstall the GPU drivers, and upgrade to either a later or earlier version. Just to rule out a driver problem on the GPU side. Secondly, any application that is not needed, kill at startup, just to remove any poorly designed driver software from the equation.

...Also, the next time you BSOD, make note of the first of the four numbers in teh parathesis; if its ALWAYS 0x00000008, then its possible you have an occasionally stuck bit at that address, hence why memtest might not be picking it up...

 

[jpsenior]

i see. (i think).

I had driver genius pro on that computer to update my drivers after i made it. Is that poorly designed/how do i know if it's poorly designed?

If i shouldn't use that software, what is your preferred way of uninstalling the gpu drivers (or any drivers from that matter). i feel like device manager never does the job or just fails. Sometimes it tells me there is no update available but the amd/nvidia/gigabyte/asus websites etc. that i've used in the past for this and other computers often says otherwise.

 

[pat]

So the computer crashed again last night. The RAM is still at 1.550v. Should i up it to 1.600v? or run some diagnostic beforehand? or a new strategy?

I would try it at 1.6v Does it seem to be more stable? And it is a computer, crash may happen from time to time.