Hardware problems / How to test system stability

[Scipio]

Hardware problems / How to test system stability

Since there are some problems with computer-software which have nothing to do with bugs or errors in the code of the program, users should also be aware of possible hardware failures.


Symptoms for hardware problems include:
- random reboots, system crahes often
- programs freeze, do not respond to user actions
- certain programs freezing during heavy use (e.g. when encoding)
- ...

Reasons for failures
- hardware defects, bad or too old parts (nothing lasts forever...)
- overheating due to insufficient cooling / overclocking
- transport damages (either during shipping or e.g. at LAN-parties)
- ...

Result: wrong calculations, parts overheating and shutting down (e.g. overheating protection on modern CPUs/mainboards), random errors or freezes, etc. etc. see symptoms above.

General CPU-information (model, temperature, ...)
- If you want to know which CPU model is in your PC, then grab CPU-Z (http://www.cpuid.com/cpuz.php#download) and in case of questions post the screenshot in the forum.
- Temperatures can be read best with "SpeedFan".
- Just download (http://www.almico.com/speedfan.php) (at the bottom of the page), install and start it.
- You can (and should) first run SpeedFan and then one of the above-mentioned stress test programs in order to watch the CPU temperature.
- The CPU temp should not rise above 65°C during 100% load; actually, it's good to stay below 60°C.



How to test system stability

1) with Prime95
(calculates prime numbers, with error detection, heats up CPU)

- Download Prime95 (http://www.mersenne.org/freesoft.htm). The download is located at "Step 1. Download the appropriate program for your OS."
- Unzip, start, then select "Just Stress Testing" during the first start of the program.
- Go to Options and run the "Torture Test".
- Let your PC run for several hours.


2) with cpuburn4 (download see below)

Quote from the Readme:
*** WARNING *** These program is designed to heavily load CPU chips. Undercooled, overclocked or otherwise weak systems may fail causing data loss (filesystem corruption) and possibly permanent damage to electronic components. Nor will it catch all flaws. *** USE AT YOUR OWN RISK ***

I've made simple batch files using cpuburn4 in conjunction with runprio.exe for maximum stressing of components.

Important: Be sure to close all open documents. If your PC crashes during testing, your data will be lost if you don't. I'd recommend not to use your PC while testing, since it will be very slow during testing (the test is run in high priority mode).

After downloading the ZIP file, do the following:
- Extract.
- Now run one of these files:

!burnP5.bat - Intel Pentium w&w/o MMX processors
!burnP6.bat - Intel PentiumPro, PentiumII&III&IV and Celeron CPUs
!burnK6.bat - AMD K6 processors
!burnK7.bat - AMD Athlon/Duron processors
!burnmmx.bat - to test cache/memory interfaces on all CPUs with MMX
!burnbx.bat - an alternate cache/memory test for Intel CPUs

- let your PC run for at least two hours (better: 8 hours, or even 24h). If it doesn't produce errors you are pretty safe that you've got a stable system.


3) Testing your memory (RAM)

- Download memtest86 (http://www.memtest86.com/) and create either a bootable Floppy or a bootable CD (see website for images).
- Boot your PC and run the test (read the instructions).

[DaveQB]

I have heard countless times that GoldMemory is a more thorough memory checker.

I am not sure if this is valid or not but an alternative

www.goldmemory.cz

Also, checking HDD for errors is also wise.

Most HDD makers have diag programs on their site for ATA and SCSI.

In fact i have a bootable CD that gives a menu to choose Goldmemory and multiple HDD testers :
Western Digital
Seagate (Seatools)
Maxtor (PowerMax)
Fujitsu
IBM
Samsung
Plus many other tools
I can host the ISO if anyone wants it, bout 23mb's

[jed]

DaveQB i wouldn't mind a copy of that ISO sounds very handy

[DaveQB]

i'll post a link tonight for everyone, when i get home from work.

[DaveQB]

Sorry for the lengthy delay, work has been busy (took my first day off in 2 weeks to watch the superbowl

Ultimate_Doom9_edition_ver1.6.iso (http://dward.us/Files/linux/iso/Ultimate_Doom9_edition_ver1.6.iso)
21.2MB

I whipped this up quick (removing retail software)

Let me know if its ok.

[jed]

Thanks for posting that, i'll try it next week as i am busy right now.

[ukb007]

Hi DaveQB

I’ve downloaded Ultimate_Doom9_edition_ver1.6.iso (21.2 MB) and burned it with NERO 6.0.0.0. The CD came out with a volume label of Multiboot Diagno and it contains a folder named isolinux which contains the following:

Basic.img (1440 KB)
bootmsg.txt (2 KB)
cdrom.img (1440 KB)
cupid.img (1440 KB)
Fuji.img (1440 KB)
gm.img (1440 KB)
ibm.img (1440 KB)
isolinux.cfg (1 KB)
kill.img (1440 KB)
Max.img (1440 KB)
memdisk (15 KB)
msnet.img (1440 KB)
NTFS.img (1440 KB)
Sam.img (1440 KB)
Seatools.img (1440 KB)
tomsrtbt.img (2880 KB)
WD.img (1440 KB)

The .txt file says:

"The Ultimate diagnostic boot CD
Creator: David Ward (DaveQB)
Ver1.6

1) Basic clean boot
2) GoldMemory 5.07

4) Seatools
5) PowerMax
6) Western Digital
7) Fujitsu
8) NTFS tools
9) Samsung

11) Tom's Linux boot
12) Network boot
13) IBM
14) Intel CPU ID tool
17) Erase disk

0) Boot from first harddisk (0x80)
a) Boot from first floppy drive (0x00)
q) Skip CD-Rom boot (try next boot device)

When no key is pressed for 300 sec, or when you hit Enter "0" will be run...

Type command at the prompt..."

I apologize for being dense. I don’t know how to use files with this .img extension. And the .txt file was helpful, but I was unable to fully understand the messages in it. Will you or someone else please tell me?

Regards

[DaveQB]

hehe no worries champ.

The disk is a 'boot' disk.

1) Put it in
2) reboot system
3) go into your BIOS/CMOS settings and make sure CD is the first boot device.
4) save and exit.
5) system will boot up off the CD with a menu (like you found in the txt file)
6) hit the number of the program you want to run. ie 1 for a clean boot etc.

the .img files are images of boot floppies.
The disk is designed so that it loads the image file intoi memory and it runs like you have booted off a floppy of said image.

Makes sense ??
Clear as mud right ?? :D

[Scipio]

Hi Dave,

to make things even clearer in your readme, you could add your explanation in your posting above to your readme file. :)

[ukb007]

Thanks, DaveQB.

As nicely said by Scipio, I'll now make a second disc with your last post in the readme.

You may perhaps make another CD-image (.iso) with this revised readme. The reason I'm asking is that not many of us mortals can read Promish, the language of the gods (programmers), even though it is supposed to be a dialect of the English language.

Regards.

[DaveQB]

Sorry everyone.

I thought it was made obvious in my first post in this thread

In fact i have a bootable CD that gives a menu to choose Goldmemory and multiple HDD testers


I will change the Readme to reflect this :)

PS i sure as hell dont consider myself a programmer, i just used Bart's way to make this bootable CD ;)

[BlackSun]

Just use the Ultimate Boot CD, this is what I use most of the time at work (computer tech)

http://www.ultimatebootcd.com/

Contains:

Hard Disk Installation
MaxBlast 3 (Maxtor)
DiscWizard 2003 (Seagate)

Hard Disk Diagnosis
Drive Fitness Test (IBM/Hitachi)
PowerMax (Maxtor/Quantum)
DLG Diagnostic (Western Digital)
Data Lifeguard (Western Digital)
SeaTools Desktop (Seagate)
SeaTools Desktop (Seagate)
Diagnostic Tool (Fujitsu)
SHDIAG (Samsung)
HUTIL (Samsung)

Hard Disk Device Management
IBM/Hitachi Feature Tool
AMSET (Maxtor)
UATA100 (Seagate)
Ultra ATA Manager (Western Digital)
SMARTUDM
ATA Password Tool

Hard Disk Wiping
AutoClave
Active@ KillDisk Free Edition
Darik's Boot and Nuke

Hard Disk Cloning
HDClone (Free Edition)
g4u

Partition Tools
Ranish Partition Manager
XFDISK (Extended FDISK)
SPFDISK (Special FDISK)
TestDisk (used to undelete partitions)
Partition Resizer
Partition Saving
Free FDISK
MBRtool
MBRWork

Boot Managers
Smart BootManager
Gujin

File Managers
DOS Navigator
File Maven
NTFS Tools
Offline NT Password & Registry Editor
Active NTFS Reader for DOS
EditBINI

Memory Diagnosis
Memtest86
Memtest86+
Windows Memory Diagnostic
DocMem RAM Diagnostic

DOS Boot Disks
Bart's Network Boot Disk
Bart's CDROM Boot Disk
FreeDOS Boot Disk

Linux-based Rescue Disks
Tom's Boot Disk
Recovery Is Possible (RIP)
BasicLinux
Trinux

System Information
AIDA16
NSSI
PC-Config
PCI

Benchmarks
System Speed Test 32

Antivirus Tools
F-Prot Antivirus for DOS (Personal use only)
McAfee Antivirus Scanner

BIOS Tools
CMOS Password Recovery

Network Tools
Freesco (free router software)

http://www.ultimatebootcd.com/

[BlackSun]

oh, and myself I use Burn In Test to test the stability of the system:
http://www.passmark.com/products/bit.htm

Memtest86 will freeze using an Athlon 64 (tried many motherboards), get the latest Memtest86+ 1.15 or better from www.memtest.org, it has been fixed. If you own an ASUS Motherboard with a VIA chipset like the A7V8X-X, memtest86+ will freeze because a bug in the BIOS. Get the standard memtest86 (not the '+' version)
Intel motherboard simply freeze using any versions of memtest, to get it working, disabled the USB Legacy Support in the BIOS.

And a good way to test your hardware is simply to do an installation of Windows XP:
if you get a BSOD during the install saying IRQL_LESS_OR_NOT_EQUAL, or BAD_FAULT_IN_NON_PAGED_AREA then check your memory using memtest.
If you get errors while the installation copy the files to the hdd, check your CD, and see if it happens each time on the same file, this is probably the CD itself. Otherwise, if it's kinda random, you better use the HDD diagnosis tool for your HDD. If you own a Maxtor and get an error code ending with S57, try a low level format (got this information from Maxtor Ireland), most of the time the problem could be fixed using a LLF.
If Windows XP can't create the partition and get you in a loop asking to create the partition again and again, same think, try the hdd diagnosis tool for you HDD, something is wrong on this side.
If you get a BSOD INACCESSIBLE_BOOT_DEVICE, make sure you have a 80 conductors IDE cable, also it's possible that you installed the wrong chipset drivers and Windows cannot access the HDD anymore. If you changed your motherboard, the HAL (Hardware Abstraction Layer) is now uncorrect for the new hardware and a new fresh installation is mandatory.

Hope this help a little ^^

edit: removed a comma in the url O_o

[techz]

Here's a few:

UBCD

http://ubcd.sourceforge.net/

Barts PE

http://www.nu2.nu/pebuilder/

[DVD_GR]

Regarding system stability issues,I would like to share my experience on these issues.prime is not needed for the purposes of video coding.My modile barton NEVER!! passed prime...and NEVER had problems with 8 hours of CCE!!maybe its 100% cpu usage but video coding doesnt really kill the cpu.Download OcBible from my sig and run some stress tests,its really helpfull.One other stability test is super pi to calculate 32M digits.They are effective for video coders.you can pass super pi and dont pass prime,and still be able to code for even 12 hours straight.if you dont pass basic pi calculations dont even try video coding!!
PS:memtest is a good program to verify that your memory,is in good shape!

[MrSeanKon]

Download OcBible from my sig Link is broken :D
Well OcBible and Guidemania are downloadable here (http://files.aoaforums.com/D171-Cpu%20Burn%20In%20Stress.html) for anyone.
The OcBible is complete package with many estimators etc...
Guidemania includes many enormous guides.

run some stress tests,its really helpfull. LOL OcBible is not a stability program.
It includes CPU Stretcher but it is not a stability program; it is a CPU burn-in program.

A complete list with must have p/c programs is here (http://forums.techpowerup.com/showthread.php?t=20251)
I voted yes for this guide I think it helps. :)

[khat17]

Some PC hardware problems could be caused by failing parts which may not always manifest themselves - or - show up while something else is being stressed and give you "false positives" kinda. A failing PSU may give you errors on any and all aspects of using you PC. I recommend periodically tearing down your PSU (don't do it if you don't know what to do or will void your warranty if you do) and oiling your fan and checking for any visibly obvious faults like burns near any component or bad capacitors.

www.badcaps.net

Get your rig cleaned out with a tin of compressed air or a compressor with a dehumidifier, or be careful if using one without it. Do that so you keep things clean and reduce the risk of stuff getting clogged up with dirt.

If you're a tech, then some tools like a POST card or PSU tester along with your troubleshooting experience would definitely help to pin-point your issue.

Too bad you can't vote twice..........or change the vote. The guide helps, but some additional stuff could be added to it by members.

[laserfan]

Too bad you can't vote twice..........or change the vote. The guide helps, but some additional stuff could be added to it by members.You do realize you resurrected a 3 year-old thread!? :devil:

The latest/greatest system stability tester is Linx, which is a Linpack gui interface.

[khat17]

I see - but it was worth it. Now badcaps and linx are in the light as well.

[brencreetle]

Hi

Ive installed Myth TV in Debian on my mac mini but I dont have any video capture hardware. Does anyone know of any suitable video capture hardware that has linux ppc drivers and will work in Myth TV? It doesnt look like any of the Elgato hardware is supported.

I dont really want to have a separate backend PC as it wouldnt look good in my living room

Thanks.

[ShinichiYao]

My system can pass every hard test(more harder more stable), but still random blackscreen and restart in an idle state.

[setarip_old]

Sounds like your system may be overheating.

Try cleaning the insides to get rid of dust bunnies, grease, etc - carefully examine and clean both your fan(s) and the heatsink...

[antonio.landon]

Use a software such as Core Temp.

A good CPU temp is around 60C..

[Amanda]

:goodpost: Do learn a lot from your post. System stability is very important as hardware failer will cause huge troubles, saying: a crash from PC hard disk always make troubles like data losing, PC black out. To pretect PC and computer data, it's necessary to maintain system stability, and preparing a backup plan can also help to avoid this kind of trouble - just using backup software free (http://www.backup-utility.com/free-backup-software.html) to copy hardware data, restoring PC in case of misfortune happening.

[finlay31]

A good routine update of softwares is a good practice for system maintenance