Slashdot Mirror
Software To Diagnose Faulty PC Hardware?
Etylowy writes "Over the years I have repaired my own PC and those belonging to family and friends many, many times. While in most cases it turned out to be restoring a system after malware/the user/Windows made a mess, or simple cases of 'follow the smell of smoke and molten plastic,' there were some nasty ones where the computer mostly works. By 'mostly,' I mean: you can boot it up, it might even work for a while, but will crash way too often to blame it all on Microsoft — what do you do then? Once you strip it of any extra hardware (which, with today's motherboards that have pretty much everything integrated, might not be an option) you are left with the CPU, motherboard, graphics card, RAM and HDD. You can test the HDD, you can run memtest86+ to check the RAM, but how do you go about testing the CPU, motherboard and graphics card trio to find which is to blame? Replacing them one by one isn't really an option. Do you know of any software that would help the way memtest helps with RAM?"
It will stress your RAM, CPU, and GPU or all at once with pretty temperature and utilization graphs (for Windows only): http://www.ocbase.com/perestroika_en/
Free means no restrictions, ironic the FSF's GPL forces restrictions, isn't it? What's your definition of free?
The phoronix test suite is a good profiler, at least it would narrow the search. But, as you observed, once you are down to the RAM and integrated devices what options do you really have expect to toss the mobo?
This is probably one of the best and most comprehensive OS agnostic boot-CD/floppy general purpose PC hardware testing and burn-in tools I've come across IMHO.
Here's its web page : http://www.eurosoft-uk.com/pc_check.htm
In any case, I recommend plugging the ATX cable into a power supply tester that presents a non-trivial load as a first step in diagnosing any PC. You'd be surprised in what ways the problems caused by out-of-spec voltages can be manifested.
jdb2
self-checking programs like Prime95 can be useful to test the computer more generally (if you've verified with memtest a failure here basically means cpu/chipset at fault).
Other things I've tried before have been (if the motherboard allows) things like significantly underclocking sections of the motherboard/processor, if an specific underclock fixes the problem you just significatnly narrowed down the list of possible failures.
there are similar programs to memtest that will check a GPUs output conforms to what it should, but if you just have random-crashy-badness that can be a pain to diagnose. Sometimes things like just running without graphics drivers for a while can help spot those problems, if the computer no longer crashes you can look a bit further away from the graphics card as most of it's capabilities won't be used.
Repairing hardware makes no sense anymore. Just swap in a new machine from the pool, so the user will be happy again, call the manufacturer to send someone onsite to replace the system board, redeploy the image, and put the machine back into the pool.
At home, i usually replace the machine before it has a chance to get old and flaky.
Well... typically you find the fault by using an application which stresses one of those components far more than any other and then seeing if the failure condition you're observing occurs more often. This is just basic troubleshooting, it's not even specific to computers.
#fuckbeta #iamslashdot #dicemustdie
Most home computer hardware failures come from "brownouts".
If you notice that your lights dim a little bit when your fridge compressor or AirCon comes on, that is a recipe for a computer failure. Spend $50 get a UPS
Btw, i noticed that my linksys wifi router was also extremely sensitive to brownouts. It would get funked up and need to be power cycled. Plug it into a UPS , no more wifi problems either.
I learned this the hard way when i moved to an old building in the east village of NYC and had 3 motherboards/cpu fail within a 3 month period.
1 1 2 3 5 8 13 21 34 55 89 144 233 377 610 987 1597 2584 4181 6765
I like the Microscope products...their newest version Microscope duo boots off of a USB stick. For machines that dont boot at all they also have a diagnostic card, its basically a pci card that has an led readout that give a series of post codes that can help diagnose if its the board, a card, memory, etc. They can be found at http://www.micro2000.com/
The handiest piece of diagnostic gear I use is actually a simple power supply tester. You would be amazed how many systems that appear to power up are actually suffering from a dead -5 or +5 rail on the powersupply. Many tend to think if the fans spinning the powersupply is ok but thats often not the case. The best part is they are cheap...around $10 for a basic one.
Hiren's BootCD contains a bunch of different utilities for doing just this. Plus it's bootable, so if you can't get into the OS you can still use the CD. It can do just about anything you'd need to in order to diagnose and repair a machine. You just gotta find it (usually the pirate bay or other torrent sites are a good place to look.)
Furmark http://www.ozone3d.net/benchmarks/fur/
Is better suited for stressing your GPU, it's also free.
When you no longer trust your CPU/motherboard, I am afraid the only option to test them would be a hardware circuit (which can make decisions using its own CPU) specifically designed for your motherboard/processor. Which I believe only manufacturer will have access to. If you are looking for a more practical solution. The only way is to eliminate the possibility of all other hardware failing (by simply removing them or using them on a good machine) and assuming it must be CPU/motherboard issue(which means you may have to junk them both and buy new ones). And dont forget to test you power supply unit (not checking it on my old PC cost me hell a lot of hours)
http://sourceforge.net/apps/trac/smartmontools/wiki is great for finding out what the drives think about their own health. Things to look out for are spin-retry counts (which lead to that annoying 2-5 seconds freeze), high reallocated sector counts (never never never use chkdsk to attempt to fix a broken hard drive. With the robustness of modern journaling file systems (HFS, extN, NTFS), storage errors are almost always hardware errors. Running chkdsk stresses the drive just as it's failing and usually pushes it over the edge -- and then users complain that you can't recover their data.
Prime 95 is a good test of CPU/RAM, as well as to see if the system remains stable under peak temperature. It's often used to burn in overclocked machines.
That's a marginal idea at best, but a common one.
While the technique of blasting a processing unit to see how it behaves at maximum temperature will sometimes find a faulty unit, many faults are not temperature related, and will not show up on this test. It's fine that you brought it up here, but something that both heats the CPU/GPU and tries to test as many pathways / as much of the instruction set as possible would be far more useful. (cf memtest86+ for RAM)
I won't join Slashcott. OTOH, If Beta goes live, I just won't be back until it's fixed. Sorry Dice.
Oh, and don't forget to check the PSU. When it acts up, it will often appear to be a hardware fault somewhere else in the machine. (often RAM, but can be MB, CPU, GPU...)
This certainly doesn't answer the posters question, but it is related and important.
I won't join Slashcott. OTOH, If Beta goes live, I just won't be back until it's fixed. Sorry Dice.
Integrated devices can typically be replaced with PCI/PCIe devices. If an integrated network or sound card gives out, it can often be easier and less expensive to shove a new device into the case and disable the old one in the Device Manager. Still, integrated devices don't go out that often. It's more common for the MB itself to go (my experience, anecdotal).
I won't join Slashcott. OTOH, If Beta goes live, I just won't be back until it's fixed. Sorry Dice.
There is no way to tell, with software, whether your PSU, CPU, or motherboard is to blame, in the overwhelming majority of cases.
It's just idiotic to say "Replacing them one by one isn't really an option". In fact, that's by far the best option. I don't run memtest for a week to find out I have bad RAM, I take 30 seconds to swap it, and find out, for certain, in no time. PSUs are equally easy to swap, AND are the more likely component to fail, so that's the best place to start.
If you don't know whether it's CPU or the MoBo, buy a new motherboard... Vastly more likely to be the cause, and pretty damn cheap just as soon as they're no longer brand new. Of course CPUs fail, but it's likely to be obvious from a visual inspection if they've been installed wrong, or otherwise abused.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
I stress my Linux boxes by telling them that if they develop a fault I'll re-image them with Vista.
Not a single one has dared to fail on me yet.
AT&ROFLMAO
It is a good start, but no more than that. Those tests are certainly not comprehensive (and should be). On the plus side, they often have your specific hardware in mind, and might possibly catch something that other tools wont. (doesn't happen often, but sometimes...)
SMART is also not the end-all of hard drive indicators. A drive can pass SMART, and still be on the way out. I've found (for those familiar with Linux) that a dd from the hd to /dev/null will often spit out errors on a drive that's getting ready to fail. A linear read is far faster than a read/write surface scan, albeit not as thorough. (can be run from knoppix live CD)
I won't join Slashcott. OTOH, If Beta goes live, I just won't be back until it's fixed. Sorry Dice.
Or you could install Gentoo. That'll eat up the CPU, RAM, and hard disk for darn near eternity.
You didn't mention the power supply.
In my experience, a "crashy machine" is almost always down to the PSU. Out of the dozens of "crashy machines" I've had to fix, only one was due to bad memory. The rest were *all* down to faulty power supplies, and all of those were due to capacitors that had failed.
I have an oscilloscope so I can easily test for ripple without needing to open up the power supply and look for the obvious signs (bulging capacitors, maybe ones that have leaked). We've had dozens of machines at work with supplies that have gone bad this way. Bad capacitors have been a real problem in recent years. Four years ago, it wasn't just in power supplies either - we had to return 70 machines to Hewlett-Packard under warranty after the capacitors on the motherboard began failing after 3 months of use. We've not seen anything on that scale on motherboards since, but we still have frequent problems with power supplies failing from "capacitor plague".
A machine of mine was actually killed by a sudden power supply failure - the PSU let the magic smoke out with a loud "bang", and there was the sound of stuff richocheting around the computer's case. That sound turned out to be bits of exploding chips on the motherboard. The only thing that survived that incident was the CD-ROM drive - all other components were destroyed.
Oolite: Elite-like game. For Mac, Linux and Windows
Swapping out is the ONLY way.
I have systems with intermittent (heat activated) dry joints on a mobo, partly duff RAM, and partly duff (rebranded at higher clock) CPU. ONLY swapping out will find it.
HTH etc
http://slashdot.org/~GuyFawkes/journal
With hardware its usually bad psu, then bad memory, then bad caps.
Then bad karma, then bad mojo.
"City hall" in German is "Rathaus" Kinda explains a few things......
My shop uses it, works pretty well. A full scan can take up to 6 or 8 hours (we set up hardware diags before leaving for the night, and in the morning on a 24-channel KVM), but it is THOROUGH. VRAM, RAM, HDD, CPU, everything is tested and thoroughly. First step should be testing the PSU, then running QT.
Of all the Universal Constants, here's one I know: Nice guys finish last
We have repaired about in excess of 50,000 machines, and I'll tell you the tools needed are very simple. The process we do is, open the machine, dust with air compressor (with humidity drier, you can pickup at sears a 4gal with drier for about $99, saves alot of money on $3-6 cans of air) and central vacuum system (a shopvac will work), then inspect the motherboard & video card for blown caps. Take off the cpu fan and inspect the compound, if it is home built, lord only knows what you'll find. Test the power supply with a digital power supply tester (one of the $12 lcd ones) if good, still open the power supply, look for blown caps. (many will have blown caps, and be causing sporadic problems the simplistic tester will not). See if machine will power on / boot. If it doesn't power on, or hangs on post, remove modem and nic if it's a seperate card, when these are blown by lightning will cause no post. Ensure the hard drive is mounted properly with 4 screws installed, less than that the vibrations will cause the drive to go bad. (don't care what operating specifications you show me, or what G-rating the drive has, this is the case) Then test memory with Memtest86+ 1.70, and the hard drive with one of the 3 versions of Seatools by seagate. (some versions will lock on some video/chipsets, if you get a long string of bad sectors on a hdd bigger than 320gb, that begin about 2/3rds way through drive, test with a different version to be sure, as there is a sector count issue with some large hds) The 3 versions are an older GUI one, the newest GUI one, and the text version. If you have even 1 bad sector - replace the drive. We do the above process on EVERY machine before we attempt to do anything else, it is well worth the couple hours it takes to do. If you make it this far, than 99% of the time, you're problem is malware/viruses. Run Combofix, look for files not removed by it, boot with Ultimate Boot CD (the WinPE based one) or something like Knoppix and manually remove them. Search the WIndows, Windows/System32, Windows/System32/Drivers directories for files created in the past month, anything suspcious is probably a malware. Rename those files. Look under Program Files, Program Files/Common, ProgramData, and Users/UserName/ApplicationData for suspicious directories and rename/delete, these are where your AlphaAntivirus, Windows Police Pro, UltimateAV, etc, like to hide. Boot back into windows, run Hi-Jack This!, remove any suspicious entries, reboot, anything left? If so, remove manually with bootcd. In add/remove programs, remove all unneccessary programs. Then run CWShredder, Malwarebytes Antimalware, Spybot, and AVG Antivirus. (Feel free to substitute legimate antimalware/antivirus tools in place of these 3, but we find these 3 work best for us. Install all Windows updates, update all sytem drivers, try browsing the internet for 2 or 3 minutes. If all seems ok, reboot one last time, and be sure you can browse the inet still. All done! This fixes pretty much everything. Other than specific issue your customer may have complained about. Also, be sure to check the amount of ram here are what we recommend, otherwise, with latest service packs, etc. machine will seem sluggish. Windows 95 - 96mb+, Windows 98/ME - 196mb+, Windows 2000 384mb+, Windows XP 640mb+, Windows Vista Home Basic 1Gb+, Windows Home Premium 2Gb+, Windows Vista Ultimate/Windows 7 4Gb+ If you don't give machine back with this amount of ram, your customer will swear machine is slower than when the brought to you, doesn't matter how untrue it is, doesn't matter how much malware you removed or how machine didn't even go into windows! CPUs/Video Card rarely go bad unless abused. Normally, your find a under-rated power supply, or defective power supply to blame. Also, if you're working with a notebook, be sure to dust the exhaust/intake vents, if still power down/lockups, you need to disassemble and recompound cpu/video chipset with Arctic Silver 5. The other thing is power problems, mouse lockups, etc many times are caused by bad batteries, try running w/o a batter installed, just ac adapter. Any battery older than 2 1/2 years old is suspect. And of course, look for broken dc power jacks.