Computing Myth #1: Software cannot damage hardware
Posted by Trixter on February 2, 2006
Oh yes it can. Here’s a story for you: In my teens, I had a friend who got the IBM 5155 (the Portable Computer — you know, the heavy luggable one with the monochrome CRT monitor inside it) from a Computerland. He was screwing around with POKE in BASIC and POKE’d a value somewhere into CGA-land, saw some pretty squiggles for about 3 seconds, then poof and we smell the familiar smell of ozone. We went back to Computerland, but they told him that there was no way that software could have damaged hardware and they weren’t going to repair it (thinking that he had dropped it or something). So my friend, with salesman watching with one eye from across the room, walks over to another one on display, takes the diskette out, boots into BASIC, writes a 1-line program, and poof, ozone, and no more monitor. Salesman didn’t quite know what had happened, so my friend walked over to another one on display and does it again, but steps aside so that the salesman can see it this time.
Let me tell you, I’ve never seen an overweight, balding, 50+yr old move so fast. He sprang like a gazelle toward us and for a second I thought I was going to get the crap beaten out of me. Thankfully, all he did was yell, about how we were going to have to pay for that, etc. This got the attention of the manager, who came out and said they’d honor the warranty and fix his machine if the kid would stop blowing up all the IBM 5155 monitors in the store.
Oldskooler Ramblings 
jim said
Yeah, and a fully-punched card could jam up card readers. Do you have any examples less than 20 years old? :)
Ipggi said
My old ATI 8500 had a bug in it with certain invalid resolutions (caused by games crashing etc) that would literally fry my expensive Panasonic 19″ CRT. Thankfully it came with a 3 year warranty, and they would cover it. But it was annoying lugging that thing to the service repairing multiple times. Especially how it was really ATI’s fault.
Trixter said
jim: No modern examples because most people today simply buy into the myths ;-)
But seriously, I don’t think there are many computing myths in 2006 because computers have become part of international culture. They’re household appliances, like televisions or dishwashers or phones. Everything (consumer) is known. So there isn’t really any magic in computers (in general) any more.
Snow_Cat said
Actually, on many Compaq laptops with the ATI chipset using interlaced mode would literaly burn the GPU.
My cm2130 for example.
Brolin Empey said
On motherboards with fans whose speed can be controlled by software (via ISA bus + I2C), software could e.g. permanently stop the fan on the CPU’s heatsink and cause the CPU to overheat. This would almost definitely have to be done deliberately, though: it seems unlikely to me that someone writing random values to random ports would write the right sequence of bytes to the Super I/O controller’s index and data ports in order to stop the fan.
You could also e.g. program a CD or DVD drive with the wrong firmware and render the drive unusable. You could not do this just by poking values in memory or I/O ports though.
Rimbo said
The initial revision of the Motorola 68000 (first chips off the line) had an HCF instruction. “Halt, then Catch Fire.” It caused an infinite loop, which caused the processor to melt.
Trixter said
HCF: not a myth, it was a test instruction left into the original run. It toggles the bus lines so fast that in some configs it burns them out.
Motherboard fans: In order for that to work, you’d also have to disable the alarm/automatic shutdown :-) otherwise the machine would promptly shut off to save itself.
ZoFreX said
I have a particularly cheap Taiwanese motherboard and although their configuration utility only allows me to throttle back the fans to 40%, speedfan will let me drop them to zero – yes, I checked, they were actually not spinning. Pretty awesome. Even with the bios temp detection, I’d be risking damage if I combined speedfan with superpi.
Zed said
Omg! This is so true! This is something that happened only a few days ago to me. I was poking around in the VGA memory area with a pointer 0xA0000000L (this is the video memory address for screen mode 0x13). I was unable to directly access the video, so I was randomly writing to the memory. And you know what happened? The picture on the screen started vibrating and poof! I heard a small bang and a flash and some smoke. This happened on my 17″ CRT LG monitor. Can you believe it, this is the latest hardware running on Windows XP, none of the two had any safety measures to stop my program written in Turbo C (16-bit) to meddle around where it isn’t supposed to. So, in fact, it is a myth that “Software cannot damage Hardware”. And trust me, I think it is pretty easy to write a virus that could blow up ur monitor; think about it.
Lesley Clarke said
We downloaded a game to two computers and within a week both motherboards had fried. I didn’t believe it but my husband said at the time that it was the game!! Any comments?
OnlineClaimsTester said
Dear Zed, please provide the system configuration, OS, the compiler used and the code that you were writing and we will replicate and test it to see what happens. Ms.Clarke, I am not sure whether a game could possibly damage a systemboard and that too 2 systemboards is kind of very ambiguous. However, we would like to investigate and see what you claim is really possible. Please specify the system configuration and the games that were installed. Also mention the kind of cooling mechanism used if any and also mention if any of the system features were overclocked…. We are very sure that every system failure has a logical reason and can be isolated.
Angel said
I think I blew the functionality of my built in microphone in my laptop by installing some drivers for the MAC OS X , you know installing that hacked version, uhm I dont know how the rest of the sound works but that thing not, and when I plug an external mic works, so I have no idea, HELPPPPP, its weird..
Unknown said
So, software cannot damage hardware…Ok, then try flashing your bios and disconnect your computer in the middle of the process and see what happens, or better yet, try flashing the BIOS with an incorrect BIOS version
Trixter said
I think you missed the point of the post. Obviously software can damage hardware… that’s why the title has “myth” in it.
The Same Random Guest As Last Time said
Reply to 5: The HCF command halts the processor and reads through all memory locations as fast as possible. It is right that it caused way too much heat for the system to work properly. The only way to stop the HCF is by toggle the RESET line or if something imporiant in the system gets fried.
(Thanks to Google for finding quotes form the BYTE Magazine, Vol 2, December 1977)
Anyway, Do anybody else know about any other undocumented OpCodes?
Me said
Might this be the cause?
http://en.wikipedia.org/wiki/Killer_poke
Brolin Empey said
Reply to 12: My MSI K7T266 Pro2 v2.0 (MS-6380) motherboard can read a BIOS image file from a floppy disk and program the image even after the user has programmed the wrong BIOS image.
My friend Andrew Williamson sent me a link to a modern example of software (accidentally) damaging hardware:
http://www.lm-sensors.org/wiki/Configurations/Sapphire/AM2RD790
Kelly said
The TRS-80 model II had a HCF instruction. The video mode could be set such that it would burn up the vide controller.
Software CAN damage hardware. « the geekery blog said
[…] Software CAN damage hardware. July 26, 2009 — aimeewilbury At least it used to be able to, as demostrated by this article: […]
tahrey said
Killing the BIOS isn’t really software killing hardware … it’s software killing software. Or in fact, if you’re pulling the power, a combination of software and (simulated) hardware failure killing the Basic Input/Output Software. (OK, the S is for System, but it’s still code. It’s still S/W).
I can completely dig the ACPI (fans etc) problem though. Certain pathological loads or poorly made heavyweight programs (this months example: Sun’s “Virtualbox” attempting to run a modern version of Linux inside of XP and 1Gb RAM) can interfere with my laptop’s system monitoring and fan control, particularly when I have a little Speedfan-like util called Notebook Hardware Control running (99.99% of the time, awesome for tweaking fan trigger temperatures and speeds, CPU multiplier, voltage, showing battery and HD details, etc). Pootling away at fairly low load, temperature in the low to mid 40s, fan hardly turning… then the heavy program kicks in… knackers the ACPI almost immediately, and before the CPU temperature has ramped up/the system monitor sensor has time to notice it. So the chip’s rattling along at maximum speed and 100% load, but the fan’s not kicking in.
Luckily it’s a laptop, and I can feel the heat starting to come thru the heatsink that’s seperated from my body by just a thin skin of plastic, and wafting gently out of the vents.
As luck would have it, there’s a portable fan-heater by my feet. Grab, set it to run the fan without heat, and direct straight inside the vent ports. By this point NHC has managed to grab a few vital cycles of CPU time and is squawking a “CPU OVER TEMPERATURE, >80’C!” message at me. With the aid of the suddenly very ironic “heater”, I manage to control this into the high 60s (still a point where I’d usually be panicking, but I know it’s survivable) until I can kill Virtualbox…. find out that the fan is now “permanently” locked to low speed because of whatever it did… restart the system, resetting ACPI on the way… and changing NHC’s settings so that the fan is running at maximum blast BEFORE restarting my doomed virtualisation attempt.
Not the first time it’s happened either… last time I had no recourse but to yank the power lead and drop the battery out. Nasty that the glitch occurs just at the point where you’re going to need it most. I’m kinda paranoid about temperatures over 60’c because a cousins laptop died, as far as I can tell, because the CPU temperature was continually up in the 60-70 celcius range (a much cheaper, crappier device running some modern equivalent of the Duron, which was hot for even a desktop chip when I had it in my old midi tower; it was very uncomfortable to use on your actual lap). I know there should be thermal trips in the hardware that will throttle, then suspend, and finally (up around 105’c) kill all power to the CPU to stop it briefly becoming a short circuit conductor (and ignition source) when the silicon’s semiconductivity is thermally broken….. but if the fan control can be taken off line in this way, I can’t trust that the other ones aren’t subvertible in some way – demonstrative video from AnandTech or otherwise.
First person to make a virus capable of causing this kind of breakdown on a global scale wins massive noteriety. Better make it quick though – the tech behind Intel’s Turbo Boost feature in the i7 is supposedly a bomb proof variant of the thermal switch.
Trixter said
I don’t know, my i7 starts to go really hot when I encode media, but then my fans hit 100% RPM at 60C and then it sort-of hovers there. And that’s stock cooling on a overclocked CPU. If there’s going to be a trojan that destroys CPUs, they might want to start with AMD…
Hitting thermal meltdown on an Atom would be a prize-worthy trick!
tahrey said
Ten years gone, I wonder if anyone ever achieved any of those things, or whether the general silence is evidence that most if not all modern machinery is essentially proof against such exploits.
Though whether Apple’s enforced aftermarket throttling of iPhones would count?
Alexandre, o Tabajara said
It can be done in MSX computers. You reverse the direction of the joystick (or keyboard?!) port on the 8255 PPI and presto, press a key and your 8255 is fried. I’ve seen that happen :o(
yuhong said
Zed: The problem is that NTVDM has to provide DOS apps direct access to video hardware for maximum compatiblity.
Brolin Empey said
While installing software to add USB Host mode capability to Maemo 5 on my Nokia N900 mobile microcomputer, I was explicitly warned about the possibility of software damaging hardware:
http://www.flickr.com/photos/brolin_empey/8076834727/
http://www.flickr.com/photos/brolin_empey/8076833405/
coconut oil acne review said
The ascending colon, due to gravity, is where most people’s digestion tends to slow down. You will be surprised at how smooth your lips feel. Date incarcerated: The vegetable oil industry and followers of a theory called The Lipid Hypothesis rounded up all natural saturated fats, butter, lard, coconut oil and palm oil, locked ’em up
and threw away the key somewhere in the early 1950’s.
Bob Pagendarm said
There WAS a student at Fresno state that wrote a program in the early 1970s to read and write on on one of the old stand alone 10 meg disk drives (2 ft sq x 4 ft high) and got the head reading at the outer side of the disc plates and writing to the inner disc plates, got the agitation ups so the entire drive fell over on its side. He was Booted from the school.
Trixter said
Shame he was booted; I would have given him extra credit. This story reminds me of the “hard drive races” some older coworkers told me about 20 years ago; in the early 70’s, the same action the student took would shuffle an unsecured unit a centimeter to one side, then the other, “walking” it across the room.
Sell Annuity Payments said
Greatest has ranked ING as A+ which can be the second greatest course of 15.
Fixed premium rates decrease annd increase with market conditions,
similar to CD rates. Even in the good war, World War
II, there was a feature that opposed the war.
http://www.youtube.com/ said
You have even more options for dry eyes; one precisely what iis named as semi-scleral gas permeable lenses.
If your eyes are medium gray, blue or hazel
provides will be hardly visible.