PDA

View Full Version : Machine_check_exception



pctek
20-05-2007, 09:08 PM
Woe.
My PC has started doing this BSOD.
I wonder if my dodgy PSU I replaced has caused some damage.

I ran memtest, it reported no problems. Interestingly though, on boot when it reports the memory it shows:

Tested: 2096128K, installed 2097152k.

But memtest was ok. PC no different with either of its 2 pieces of ram in.

It BSODed during FEAR.
It BSODed when I was in DXDiag, just clicked Display tab.
BSODed when I was part way into starting the latest ATI driver upgrade, its setup part.


Put my X300 in now (sob), ran FEAR again (lame settings admittedly) and no crashes yet.

Not entirely convinced its the card yet though..
Any second opinions?

CPU is not overclocked and has never been. Its a AMD x2 5000+. New.

Ram is 2 x 1Gb Kington DDR2-667. New.

Graphics - the suspected one: X1900XTX, 14 months old.

pctek
21-05-2007, 07:45 AM
bump

The_End_Of_Reality
21-05-2007, 07:49 AM
How much RAM does Windows pick up?

Did it usually show those amounts?

What is the exact BSOD error? Any drivers/files in that error?

Try tried the card in another PC?

pctek
21-05-2007, 08:26 AM
2Gb.

Can't remember if the startup amount was always showing that.

STOP: 0x0000009C

No drivers

Nope. But I have put mu X300 in and the error hasn't recurred.

I googled it to death.
In a few forums there are people saying they got it after PSU problems.
Causes varied hugely, CPU, RAM, motherboard, graphics, heat, PSU.

The_End_Of_Reality
21-05-2007, 08:36 AM
How long does FEAR run for before crashing?

It definately looks to be GFX related... (I take it that you can click the display tab of DXDiag with the x300?)

The amount of installed RAM is fine, the tested RAM is 1MB less than installed, wait up... now that I think about it, I have that too... I have never had a problem with it. So that is fine.

I think it is likely that the PSU has damaged the card, possibly with a power spike on the rail that damaged the 3D side of the card...

pctek
21-05-2007, 10:55 AM
How long does FEAR run for before crashing?

It definately looks to be GFX related... (I take it that you can click the display tab of DXDiag with the x300?)

The amount of installed RAM is fine, the tested RAM is 1MB less than installed, wait up... now that I think about it, I have that too... I have never had a problem with it. So that is fine.

I think it is likely that the PSU has damaged the card, possibly with a power spike on the rail that damaged the 3D side of the card...

It didn't go that long at all. Crashed on a fight scene. The second time I loaded it up and went into the settings and was changing Shaders and it crashed then.

Oh? Why does it do that, I can't remember seeing that with RAM before....then again I don't exactly watch that bit, its usually too quick.

Yeah I've played FEAR thismorning with the X300, cranked the settings up to where it started to give me warnings about performance but no BSOD since.
Messed about with the cubes in DXDiag, ran 3DMark, various of its tests. Zilch.

&&^%$ SILverstone PSU!!!!

SolMiester
21-05-2007, 11:17 AM
Woe.
My PC has started doing this BSOD.
I wonder if my dodgy PSU I replaced has caused some damage.

I ran memtest, it reported no problems. Interestingly though, on boot when it reports the memory it shows:

Tested: 2096128K, installed 2097152k.

But memtest was ok. PC no different with either of its 2 pieces of ram in.

It BSODed during FEAR.
It BSODed when I was in DXDiag, just clicked Display tab.
BSODed when I was part way into starting the latest ATI driver upgrade, its setup part.


Put my X300 in now (sob), ran FEAR again (lame settings admittedly) and no crashes yet.

Not entirely convinced its the card yet though..
Any second opinions?

CPU is not overclocked and has never been. Its a AMD x2 5000+. New.

Ram is 2 x 1Gb Kington DDR2-667. New.

Graphics - the suspected one: X1900XTX, 14 months old.

Hi PCtek, what psu is that, the 1900xtx are beasts for power!

edit...didnt see silverstone psu, ok then, only happens with that card......?!

Suppose the card could be fried, or the psu amperage on the 12v+ not enough, what model psu?

The_End_Of_Reality
21-05-2007, 11:22 AM
I am not sure why it does that, if I run Memtest it shows the amount of RAM being tested is lower (by only 1MB or so...) than the actual RAM installed, I am not sure why, possibly due to loading the BIOS into RAM? (I have no idea just a guess)

dugimodo
21-05-2007, 01:19 PM
yeah what's the replacement psu? does it have sufficient grunt.

if the graphics card has 2 power connectors are they on the same rail or seperate ones? I read a review of my 600W silverstone where they had fixed a power problem by using a molex -> pci-e adaptor to spread the graphics load over 2 rails because the manufacturer had stupidly put them all on one rail and it couldn't take the load. ( dual ATI cards in crossfire mode )

pctek
21-05-2007, 03:17 PM
Enermax 620W Liberty Modular Power Supply Model: ELT620AWT.
12v info only:
+12V1 0.5A/22A, +12V2 0.5A/22A, 432W( 36A )

Its not this PSU, its been in a while now.


The previous Silverstone OP650 would cause the PC to shut down at random. I sent it back and got this instead.

I previously had an old Enermax 485W which gave no problems.

pctek
21-05-2007, 03:19 PM
Memtest it shows the amount of RAM being tested is lower (by only 1MB or so...)

Er.........but I meant it shows this on the POST screen.

The_End_Of_Reality
21-05-2007, 04:29 PM
Er.........but I meant it shows this on the POST screen. Yes, I know, mine does too, I was using Memtest as an example :rolleyes:

Graham L
23-05-2007, 03:11 PM
It is sensible to not "test" the bottom 1MB of memory in the boot sequence. The BIOS is ROM; some of the 1M space might not be there, so looking for it would be disastrous. The memory sizing operation checks whether each memory block is present by writing to a location at the start of the block and reading it back. If the value is the same, it steps to the next block. If it doesn't get the same value back, it stops. That might make your GBs of memory be reported as 1MB, or less. :D

pctek
23-05-2007, 06:41 PM
It is sensible to not "test" the bottom 1MB of memory in the boot sequence. The BIOS is ROM; some of the 1M space might not be there, so looking for it would be disastrous.

? Not there? Where is it then?


Anyway, my graphics card has been replaced and the new one is on the way back to me.
So I guess it was graphics..................Morningstars techs are pretty thorough.

Graham L
23-05-2007, 06:51 PM
Remember that the lowest 1 MB of address space is that specified for the IBM PC in 1982. The BIOS is at the top of that. There's space allocated below that where extra ROM BIOS extensions can be put (E000 for VGA video?), other space for NICs, and RAM starts at 0:0 and goes up from there. So there will usually be holes where no memory exists. The early AT POST would count up from 0:0, report conventional memory, then look for any extended memory and report that. I don't know what you would see if you managed to put 640 kB of memory on a new motherboard: maybe the POST would report that much. But I'd guess it would say there was none. If there are GBs or lots of MBs, it's easiest for the sizing routine to just ignore the lowest 1 MB of address space. With 1GB that's 0.1%. ;)

And, I'll repeat, it's not a memory "test". It's just a check for presence of r/w memory.

pctek
24-05-2007, 08:17 AM
And, I'll repeat, it's not a memory "test". It's just a check for presence of r/w memory.

Its interesting because this doesn't show on my neighbours PC, its just reports the lot.
And a customers old HP which had another PC133 256mb added to it, would report a total of 370. Until I replaced the faulty module.

Graham L
24-05-2007, 04:47 PM
Maybe some BIOS coders are honest. ;) That first megabyte is almost certainly not used by Windows, because any RAM in the high end is overlaid by the ROM containing the BIOS code. The very low end is probably still used by the BIOS tables for DOS. Look at what's in 0040:0000 and up from there for the COM: and LPT addresses. They're probably still put there. ;)

pctek
24-05-2007, 06:49 PM
Well I've had a lot of weirdness here in the last week.

We had some friends down from Auckland and the following occurred during their stay:

BSOD - Graphics card, as I have already said, probable cause the dodgy PSU I had in before.

Dishwasher flooded. It has always leaked a tiny bit, this was a full flood. Only did it the once. No reason we could see.

Dryer decided to not heat, just spin cold. Only the once. Next day it was fine.

TV, DVD, Sky, heater all plugged into multipower thingy. The lot shut off. No, there wasn't an overload cutout on the power thing. All other electrics in house fine. PC was running in another room OK. Circuit breaker fine. As we stood pondering, it all came back on.

They departed, leaving us with a Rhinovirus and some appliance paranoia.

Graham L
24-05-2007, 07:00 PM
Just wait until you reap the full benefits of the IP6 extended IP addressing scheme when every electrical device has its own IP address, and can be managed and have its software automatically updared through the Internet. (Of course, to have an Internet address, every device will have to have a computer, and an OS ... even a manual toothbrush).

This will obviously be a very profitable state of affairs, so there's really only one corporation in the world with the experience and skills necessary to provide the software and manage things like bugs and security patches in a totally reliable, errorfree, and, most important of all, secure way. When needed, I'm sure Bill will volunteer to do it.

zqwerty
24-05-2007, 08:31 PM
Spheres of discontinuity (SoD) float through the Universe, sometimes they are small (as a dishwasher) other times they are larger than the Earth. A Sod cause all the known laws of Physics to temporarily be invalid and confusion rules, the condition is not permanent.

pctek
25-05-2007, 09:06 AM
:lol: