Random PC restarts/errors - please advise

[EstoyLoco]

Hi all.

I just finished my new build. I am using

Gigabyte GA-MA69G-S3H using lastest F2 bios
Antec 500W NeoHE PSU
Sapphire X1950 Pro
AMD X2 4800+
PQI 2GB DDR2 PC6400 DUAL Channel/ MAD42GUOE-X2
Seagate Barracuda 7200.10 320GB 7200 RPM HD
ASUS DRW-1814BL
Optiarc DVD RW AD-7170A

I used Paragon HD manager to move around HD size after creating 2 partitions. C and D drive. Windows event manager kept reporting bad blocks on drive D.
Ran CHKDSK /r it cleaned out invalid security links but reported NO bad sectors.
Ran CHKDSK a few times again and reports came back no errors.

Ran Seagate SeaTools, with Long Drive Test HD came back successful.

Ran Memtest86 for over 7 hours and came back successful ran both versions Memtest86 v3.3 and Memtest86+ 1.7.
Also ran Windows Memory Diagnostic (extended tests) and it came back all successful.

I can run 3DMark just fine without crashing.

However, when I run Prime95 it always fail randomly with expected rounding error etc. It is never at the same spot it is just random.
Setting dual affinities for dual core and memory setting 895MB each instance of prime.
I've had Prime run anywhere from 5 mins to 3 hours max then failure occurs.

The errors I see is, sometimes my IE will just close byitself. When I click on a link, it just closes window and gives windows error report tool.

The times I've noticed it restart is usually when I am running Windows Media Player. There have been other times when it has restarted by itself but I cannot remember what I was running. The majority is when WMP is loading or I am changing files in playlist.

I cannot figure out what is causing these crashes, input would be greatly appreciated.

Thanks in advance.

 

[montag451]

That points to cpu prob.

Might not be RMA problem - might just need to tweak voltages - go to cpu forum and ask around.


BUT on the other hand, could be a power issue as well - Have you got a spare psu lying around?
What are your cpu temps in BIOS?

 

[cprince]

It could be memory issue too. Check the voltage, speed, and timings. Memtest86 does not catch everything.

 

[montag451]

Ok,
Lets sum up.
Could be one of the following:
cpu
psu
temps
voltages
RAM
and while we are at it, mobo.

 

[EstoyLoco]

My system temp is ~40-41 C idle
CPU ~43-44 C idle

10 Minutes of Prime holds temps steady at:
40-41 System
55-56 CPU

I am not sure if these temps are good or not.

My motherboard is reporting the following voltages:
VCoreA 1.300V
+3.3V 3.230V
+12V 11.730V
CPU Vid 1.300V

Am I running Prime correctly?
I am using 2 instances with Torture Test setting
MinFFT 8, MaxFFT 4096, Run FFTs in-place UNCHECKED, Memory to use 895MB, Time to run each FFT size 15mins

Unfortunately I do not have spare memory/psu ;(
BTW I am running everything at stock settings.

 

[montag451]

Can you go into event viewer (ctrol panel - admin tools).
Note and red/white crosses.
Post back with the source, Event ID, and the full text description.

 

[EstoyLoco]

Update: Running 2 instances of Prime95, setting at 890MB (vice 895MB before), 1 Hour, results: no errors so far, temperature holding stead 41 system / 55 cpu.

The random restart does not leave a event log.
The restart isn't too often the last I recall was 7/1 that was when before I ran CHKDSK /r and then repartitioned drive D with windows setup cd and did full format.
And it was when I accessed Windows Media Player. I am not sure if my DVD-RW drive is affecting it, since I was burning at the same time and then ran WMP also.
My DVD-RW Optiarc is using a modded firmware.

Actually NM I know it has restarted since the format, I just dont quite remember date.

Here is last 3 crashes (program closing out by itself):

Date: 7/7/2007
Source: Application Error
Event ID: 1000

Faulting application iexplore.exe, version 7.0.6000.16473, faulting module mshtml.dll, version 7.0.6000.16481, fault address 0x000972e5.

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

------------------------------------------------------------------

Date: 7/7/2007
Source: Application Error
Event ID: 1000

Faulting application iexplore.exe, version 7.0.6000.16473, faulting module mshtml.dll, version 7.0.6000.16481, fault address 0x000d7b6a.

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

------------------------------------------------------------------

Date: 7/5/2007
Source: Application Error
Event ID: 1000

Faulting application iexplore.exe, version 7.0.6000.16473, faulting module ntdll.dll, version 5.1.2600.2180, fault address 0x00002fa5.

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

------------------------------------------------------------------

Date: 7/5/2007
Source: Application Error
Event ID: 1000

Faulting application iexplore.exe, version 7.0.6000.16473, faulting module mshtml.dll, version 7.0.6000.16481, fault address 0x000bf1b4.

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

 

[EstoyLoco]

Ran Orthos (switched from Prime didn't know it was dual core ready) and this was the result:

2:Test 3, 340000 Lucas-Lehmer iterations of M331681 using 16K FFT length.
1:Test 3, 340000 Lucas-Lehmer iterations of M331681 using 16K FFT length.
2:FATAL ERROR: Rounding was 0.5, expected less than 0.4
2:Hardware failure detected, consult stress.txt file.
2:Torture Test ran 2 hours, 36 minutes 32 seconds - 1 errors, 0 warnings.
2:Execution halted.

1:Torture Test ran 2 hours, 36 minutes 32 seconds - 0 errors, 0 warnings.
1:Execution halted.

Looks like it was on Core 2.

Anyways, I upped my Memory voltage by +.05 and rerunning the stress test.

 

[EstoyLoco]

Well I solved the problem.
Quite disappointing actually, I guess I will take this as lesson learned for buying cheap value ram.

After reading forums on the memory and how the particular one people were getting memtest errors with default voltage. I upped my voltage on my memory.
It started to improve the Orthos stress test, each voltage amount that I bumped up would make the test longer.
Ran Orthos 9 Hours @ 2.1V, thats +.3 from the default voltage >.<
The PQI memory line for my memory states 1.8V.

Quite disappointing that I have to run the memory at such high voltage just for stock settings.
Actually the memory isn't even running at stock 400MHZ.
CPU-Z reports 358MHz. x2 4800 (2500Mhz / 7) ~= 358

So that means my memory has a hard time running 358Mhz @ 5-5-5-18 at 1.8V
how crappy is that?

Or is that normal for DDR2 800 to be like this.

 

[noobish]

bad ram set? lord knows, could be little bug runs loose around.

im using ddr 800 stock @ 5-5-5-15 OCZ SLI and i tighten down to 4-4-4-12 and it seem fine. Its stock voltage preOC and after.

 

[simondedalus]

Usually this can be caused by a low power PSU but you have a 500w so that should be sufficient.

I suspect its a memory problem - you said you have not ran memtest86 (http://www.memtest.org) yet, so try that. Its quite amazing what kinds of problems bad memory can cause...

 

[simondedalus]

I had the same sort of problem, and I think it was due to my XP install being corrupted. I put the XP CD back in the cd rom and repaired my existing install of XP. This took care of the problem, and I'm still not sure exactly why it started. Anyway give that a try.
I think the options I was going to just format and reinstall, but when I went to do that it checked for existing versions of Microsoft OS and if found my XP install that was corrupted, and asked if I wanted to repair it by pressing 'R' So I did and all is well now. I hope this helps in a way.

 

[EstoyLoco]

"Ran Memtest86 for over 7 hours and came back successful ran both versions Memtest86 v3.3 and Memtest86+ 1.7. "

It is the memory my computer has been running stable since I either: dropped my Memory speed down to DDR2 667 @ 1.8 V or run it at DDR2 800 @ 2.1 V

I guess that will teach me a lesson about buying value line memory.