Stability Problem

FuryRosewood

Active Member
Okay i have been living with this problem for a year, and cant really trace it down, i just looked in my windows event viewer and came up with this critical event:

A fatal hardware error has occurred.

Component: AMD Northbridge
Error Source: Machine Check Exception
Error Type: HyperTransport Watchdog Timeout Error
Processor ID: 0

The details view of this entry contains further information.

EventID 20
Version 0
Level 2
Task 0
Opcode 0
Keywords 0x8000000000000000

EventData

ErrorSource 3
ApicId 0
MCABank 4
MciStat 0xfa00000000070f0f
MciAddr 0x0
MciMisc 0xc00a0fff01000000
ErrorType 7
Length 928

- This error is really infuriating me as it halts the computer without throwing a bluescreen 100 percent of the time, i dont get any reboots, i simply get a stop error, its just frustrating as hell. I want to wait for bulldozer to get a new board for this thing, because i happen to believe this shitty biostar board is the culprit...
 
Well as a very inexperienced opinion presents, try increasing the NB voltage bit by bit. Since the given component is the northbridge for the failure, if you're OC'ing by the FSB at all then it is a valid issue. Increasing the NB voltage may hlp and if its a mobo issue then a mobo issue will be solved by (obviously) a new mobo. Hope this helps, seemingly random bluescreens are the most frustrating.
 
its dead stock, so i dont see northbridge being a problem, ill try memtest, but i have a feeling even if i run for 48 hours i may find nothing, i have a sinking feeling its this biostar piece of crap :/
 
three hours into the test, no errors, i reseated the sticks before i started to be certain nothing was amiss. will come back after its done 7 hours. im fairly certain there is no memory problem here
 
its dead stock, so i dont see northbridge being a problem, ill try memtest, but i have a feeling even if i run for 48 hours i may find nothing, i have a sinking feeling its this biostar piece of crap :/

how is it stock? the processor you have listed in you sig the AMD Phenom II X4 955? Thats a 3.2ghz processor not 3.6, so you have it overclocked correct?
 
sig was what it was clocked at, when i ran into issues, i put EVERYTHING to stock prefrences, hold the crucial ballistix. it is running at looser timings. 5-5-5-15 @ 1.8V which is still perfectly stable. system atm is running at 3.2 ghz, its a C2 stepping phenom ii 955 but thats not important. currently memtest x86 has been running for 9 hours, it passed 12 times with 0 errors. this is NOT a memory problem. i need some advice on what to do about the northbridge, voltage settings wise.

http://dl.dropbox.com/u/14630925/p_00152.jpg

-yes i know the timings look stupid, and that makes no sense, its ddr2 800 at 5-5-5-15 timings, disregard the stupid output. its been running for 9 hours and has not found anything. something else is going on here.

-i just bumped the northbridge voltage up by 0.05v, im not sure that will make a whole hell of alot of good, but im trying something tiny as i wait, voltage to the memory has been restored to 2.0V as recommended by crucial, and their running at 4-4-4-12 as that never threw any errors either.
 
Last edited:
It says AMD Northbridge, which would be the processors memory controller. And it says HyperTransport Watchdog Timeout Error which is the bus it uses to the memory. Look and see if your HT and CPU/NB voltages are at default.
 
Thats the stupid thing, it doesnt really show me the voltages, tho i may look into system health real fast and determine it. The bios on this thing is really hit and miss... hell on the splash screen they spelled version wrong... -_-

i do know the HT speed was 2000mhz at default, and the i am not sure what the voltage is, i am checking now.

I just grabbed the readings from the system, let me put them in a sensible fashion here.

CPU Speed: 3200Mhz
CPU Volt: 1.376V
Chip (Set?): 1.136V
HT Link Speed: 2000Mhz
HT Voltage: 1.168V
DDR2 Speed: 800Mhz @ 4-4-4-12 (yes i restored to the ballistix factory recommended settings)
DDR2 Volt: 1.968V
5USB Volt: 5.206V
+3.3: 3.312V
+5.00: 5.026V
+12.00: 11.968V

-defaults for the NB Voltage were 1.125, i bumped it up one increment to 1.150
-default for the CPU voltage was 1.35, i left it at that, it never gave me a issue at the stock speed at that voltage as far as i can tell, tho i can bump it to 1.375 if needed.
 
Last edited:
I know some C2 had a hard time with 4 sticks of memory, but usually that was with 1066. Take two sticks out and see if the problem goes away or most of the time just bumping up the CPU/NB would fix the problem. Just a thought.
 
ill do a little at a time, until it gets to a point that im stable, also fighting compiling problems in Secondlife...dumbass devs commiting shit to the tree that makes life miserable for everyone... *grumble grumble*

ill try bumping the voltage a little on the cpu, i just am wary of how much i add, as the voltage bumps make a bit of heat in the cpu, and its become hot as it is, even with the h50...it was just a stretch fitting that damn radiator in the 1200, as stupid as that sounds, the darn resivoirs on the ends of the rad make it hard to line up against the fans at the back, i need to figure a better way of mounting it, looks like i could put it against one of the harddrive bays where they have the little extra fan mount for the spare gpu pusher fan, but i need to look more into it. those tubes to the pump are not very long and i need to be able to get the filters out of the bays, they clog quick in my house, two three days tops..and their clogged...its rediculous
 
every cpu has its own boundaries, hence why my cpu is running 3.6 ghz, if i go any higher i cannot get C&Q to function, which is strange but i can live with that.

-boomed again, just bumped the vcore to 1.375 and nb to 1.1875, but then finding the latency was getting HUGE with the vcore increases, i bumped back to 1.35 on the vcore, i just cant take that much hesitation with a vcore increase, thats why hitting 3.8ghz was annoying, i had to pump this thing to 1.425 and it was slow as hell to load...
 
Last edited:
after another series of crashes, ive started running OCCT and started getting cpu errors on all cores. i have tried boosting vcore however it still fails the test after less than 20 seconds, i believe the cpu is toast. im still unsure about the motherboard in the same case, i am just frustrated, i want a stable machine, its making me wish i just bought a dell or something...this is pissing me off for how much i spent on the hardware involved in this...
 
i dont think its the northbridge, i think its the chip at this point. i think im going to file for a RMA and see what happens, if it comes back and still isnt working, im just going to change boards and ram, this biostar board has been trouble since day one, i need something that wont waffle when you look at it funny :/
 
yeah it might be a bad cpu or hell liek you said a bad board, im running that same cpu, on a msi board, and im overclocked to 4.2ghz no problem. you might just have go a bad batch of hardware
 
i think im going to file for a RMA and see what happens, if it comes back and still isnt working, im just going to change boards and ram, this biostar board has been trouble since day one, i need something that wont waffle when you look at it funny :/

Good idea, dont say anything about overclocking to AMD.
 
Back
Top