Strange Decompression Problem

[Colonel]

One of my rigs (Pentium D 925 @ 3GHz, 2 GBs DDR-II @ 667 and a WD Rapter on an Asus Pundit P1-PH1), seems to be incapable of decompressing any zip/cab/rar/ace/etc. files. 7Zip, WinRAR, the OS built in 'ziped folder' support, 'Windows Installer' CABs, all seem to be corrupt.

However, if I decompress the same files on another machine and transfer the flat files over the network, the sick box can execute them just fine.

Windows installed and runs fine. I have tried XP SP2, Server 2003 R2, Server 2003 R2 x64 and even Longhorn Server Beta 3 x64, and they ALL experience the exact same symptoms. I used all the latest drivers from the manufacturers' sites, and flashed the BIOS to the latest 'stable' release.

I've run a boot-able Memtest86+ for over 8 hours with no errors.

Prime95 on the other hand was kind of strange:

• A 'Small' 'Torture Test' failed in less than a minute.
  
• A 'Blend' test, forced to 'Core 0' ran fine for 8 hours.
  
• A 'Blend' test, forced to 'Core 1' has been running fine for about an hour, and is still going.
  

Any one had any similar problems or recommendations on what to test next?

Thanks

Edited July 6, 2007 by Colonel

[bonestonne]

if you've got plenty of time and a spare drive, you could install linux and see if its able to decompress the files. theres .RAR support for linux if you check the synaptic package manager, which is how you install most of every program you need. not entirely sure about 7zip though, i've never even had a 7zip file, so i can't give you much info on that. if it fails in linux, then theres definately something wrong somewhere...

more importantly, if you open up speedfan and you look towards the bottom where it has all the voltages, what are each of them reading? if you don't have enough voltage, that could be a problem.

[Colonel]

Ok, I think something is terribly wrong. Middle of the night last night, it started a solid beep. Now, it wont post (still solid-beeping).

[nitroshift]

What BIOS is that? I mean who`s the manufacturer? If it's AWARD BIOS, that sounds like a RAM issue according to this: http://www.computerhope.com/beep.htm#04

[Colonel]

I pulled stick two, re-seated stick one, and it was fine.

Added 2nd one back in, solid beep.

Pulled stick one, moved stick two to slot one and it was fine.

Added stick one to slot two, solid beep.

Back to just stick one, reset bios defaults, try both sticks, solid beep.

I am using two identical IGb sticks of Kingston Value RAM (KVR667D2N5/1G). The mobo is an Asus P5R8L, with an AwardBIOS. I flashed it to v0801 from the Asus site a few months ago. I also noticed that the FSB was only 200 MHz, not 800 MHz, but the BIOS sucessfully identified the CPU.

How does a DIMM go bad durring normal operation? I mean, the system was on and running when the first solid beep started, and I was asleep so it was idling. The system is behind an APU UPS, wich repoted no problems around the time it happened.

[nitroshift]

From what you`re saying it seems that the RAM sticks are OK, the issue is with the RAM slots on the mobo.

[nmX.Memnoch]

I also noticed that the FSB was only 200 MHz, not 800 MHz, but the BIOS sucessfully identified the CPU.

Intel CPUs from the P4 up use what's called a "quad pumped" bus. This basically means that 4 times as much data is processed per CPU cycle then on older CPUs. 200MHz actual x 4 = 800MHz effective

Your problem could be any number of things. If you have other systems with parts that will work on that motherboard you can start there (i.e. if the CPU from the other stable system will work, try that...same with the RAM, etc). At first I wanted to say that it sounds like the cache in the CPU was going bad. This can happen if it runs too hot for too long. But after you mentioned swapping the DIMMs around and getting varying results, it could be the memory controller in the north bridge (which would mean a motherboard replacement).

[Colonel]

OK, so i just got a 0x0000009C (Machine Check Exception) BSOD... Something is very wrong.

Based on the history, any thoughts on what piece of hardware is to blame?

[nitroshift]

OK, so i just got a 0x0000009C (Machine Check Exception) BSOD... Something is very wrong.

Based on the history, any thoughts on what piece of hardware is to blame?

At this point you should configure your machine for a complete memory dump, archive the .dmp file, upload it somewhere on the internet and come back with the link to download it. Give cluberti some time to have a look at it and he will tell you where the problem lies.

[Colonel]

Ok, so this box is a server in my test lab, running as a domain controller. The system boots, loads windows, and the logon screen's animated bar freezes durring "Active Directory is starting...". After several seconds of frozen-ness I get the BSOD.

Once the system BSODs, it sits at "Beginning dump of physical memory", and the HDD light stays off. Since my last post, I have been able to boot to safe mode (and verify the dump settings), but have been unable to boot normally.

Edited July 6, 2007 by Colonel

[puntoMX]

I thought I drop in...

You say a solid beep, that´s the CPU error.

If you have a second CPU, replace the "bad" CPU and see what it does.

If it doesn´t work try an other PSU.

If that doesn´t work your Mobo is probebly "bad".

If that wasn´t the problem, then I should be ashamed of my self .

[Colonel]

Ok, so I formatted, installed XP Pro SP2 and followed cluberti's Creating memory dumps post. The Ctrl + Scroll caused a BSOD in testing, but once the system froze I was unable reproduce the BSOD, so I can't get a useful dump.

According to the BIOS hardware monitor all the voltages are ok, and the CPU is running at ~43C.

Unfortunetly I don't have any spare CPUs or PSUs for this box. Any other ideas on how I can narrow this down?

[puntoMX]

If you bought the part together then take it back to the shop I would say...

[bonestonne]

how many DIMM slots are there? my friend had a similar problem after [quite stupidly] putting a DIMM in backwards and then quite literally burning the board.

check the DIMM slots on the board for damage, check both DIMMs for missing pins or contacts.

if both work individually, then you may be looking at a mobo problem.

[Colonel]

The board has 2 DIMMs. The 1st one is required, as it is used for shared video memory. I have tried multiple chips in the first DIMM, with no luck. If I add a second chip to the second DIMM, I get the above mentioned solid, never-ending beep.

Guess I'll need to find a PCI video card so I can disable the integrated video.