Pushing a CPU to Heat Death, Intentionally 291
sdougal writes "This site is showing a Pico-ITX board running Ubuntu with no cooling whatsoever. They even let the public guess how long it would last: 'Last week thousands of you placed bets on how long the new Pico-ITX board from VIA, the VIA EPIA PX5000EG, can last without any cooling whatsoever. An ARTiGO Builder Kit was offered as the grand prize. Yesterday afternoon the voting stopped and the Naked Pico Challenge started in earnest. We simply loaded up Ubuntu 8.04, set it to work playing an mpeg-4 video and then removed the heatsink, leaving the CPU and VX700 chipset bare to the world. We recorded the event here in this video and set up a live video stream so you punters can keep a watchful eye on the PX5000EG as it works away.'"
Ehh, it's been done before (Score:5, Interesting)
It was an expensive lesson in the importance of the heatsink.
Of course, many of us can remember back when CPU's didn't even need heatsinks. My first build was a 486SX with a zif chip slot and no CPU cooling--hard to believe now.
The rule of thumb is.... (Score:5, Interesting)
A CPU chip with 900+ pins run a bit cooler as it's a it more than one square inch if you an include the substrate, and a certain percentage of the heat will conduct itself down the pins.
Ummm, copying VW folks? (Score:4, Interesting)
Yet again, "on the internet" somehow makes it original...
Re:Ehh, it's been done before (Score:5, Interesting)
I once sent an amd k8 system to a friend in the mail. I made the mistake of leaving the big heatsink (I think it was a barton chip and those were VERY hot back in the day) attached. the pc was sent ground, I think, and so it didn't get the best treatment. turns out that the heatsink came off the cpu socket and was doing some kind of 'round the world tour' inside the pc case! when he opened it up, there were ding marks from the sharp edges of the heatsink all over the mobo
that was bad. but it gets worse. my 'genius friend' decided to just try it as it was and not even bother to fix the heatsink back to the chip!
I think in 5-10 secs, he -guaranteed- that that system will never run again. I would have liked to know if the mobo was still working - but now, the whole thing is toast.
he didn't know? really? a BIG HUGE HONKING heatsink and he thinks he can turn on a system without it?
sheesh.
now, that was years ago. today with the core2 arch, you almost don't NEED a heatsink. its amazing. I have overclocked core2 chips (see 'BSEL mod' for changing 800fsb to 1066fsb via some conductive paint) and STILL the chip is cold to the touch when I run memtest86. my bsel mod photos are here, btw: http://www.flickr.com/search/?q=bsel&w=47907743%40N00 [flickr.com]
its now my usual procedure to install a fan speed control and set it to MIN for all my core2 systems that I build. I love the fact that even at slowest rpm, it still never gets hot enough to even pull your hand away from the hs/fan. amazing..
I also do have a via epia that I use for my mythtv box:
http://farm3.static.flickr.com/2010/1890660635_273662e3c9_o.jpg [flickr.com]
http://farm3.static.flickr.com/2299/2005750966_a1b8d242b3_o.jpg [flickr.com]
in that 2nd photo, you can see its drawing 24watts (with a kill-a-watt lcd meter). its 100% fanless, uses a 1ghz cpu but it DOES get quite hot to the touch so I leave the top case skin off; that way I can get by with no fan at all. its been doing my myth-tv recording (using hdhomerun HD tuner box, networked) for about half a year now; no reboots and very reliable.
low power systems are cool
AMDs don't need CPU fans, either (Score:5, Interesting)
Like the ole Timex watch that "took a licking and kept on ticking" my desktop box, an ancient AMD Sempron 2600+ with a VIA chipset, unknown to me, lost its power connector to the CPU fan, which I only discovered by accident when replacing a hard disk drive. The CPU was hot enough to scald my finger, but neither its performance nor its stability has suffered one bit.
Of course, the heatsink was still connected. But the Sempron was IIRC most definitely NOT a low-power cpu.
Yes, I reconnected the CPU fan. But at least I know my sh*t can take the heat.
No video is available ;o{ .
"Heat Death" (Score:5, Interesting)
Done, accidently, before (Score:4, Interesting)
It was some ancient AMD chip that we literally couldn't buy new fans for any more, so we just snipped the speaker cable and let it carry on.
Naturally, the Linux guys claimed if it had been Windows, we'd be looking at a dead server at this point in time
My experience with no heat sink (Score:5, Interesting)
/var/log/messages: (Score:4, Interesting)
If I re-encode a movie I get: Do I care? Not really. Been like that for 3 years now. When it dies I'll swap it for a less powerful CPU and go totally silent.
Even with a heat sink (Score:2, Interesting)
Re:Open != better cooling (Score:3, Interesting)
Not a challenge... (Score:3, Interesting)
http://www.genesi-usa.com/efika.php [genesi-usa.com] - plug plug
That system runs at 1W@400MHz, although has no video-accelerating northbridge to add to the heat, it can play that MPEG4 video just fine (I am playing something similar now). We've designed it so the 2.5" hard disk actually sits about 5mm from the top of the CPU - if anything we're making cooling harder, and there is NO heatsink. The CPU does NOT power manage into SpeedStep style states - it just runs at 400MHz or "standby" (where it cannot run code until an external interrupt).
It runs fine. Mine's been on 24/7 for nearly a year, barring moving it around and connecting it up to things like new hard disks, changing power strips or measuring the power it uses. It never overheats.
What's the challenge meant to be? Just how crappy Via's chip needs to be that it CAN'T run at 500MHz on a 90nm process, and do without a heatsink of some kind?
Re:Ehh, it's been done before (Score:2, Interesting)
Re:Ehh, it's been done before (Score:1, Interesting)
Fortunately I had a race car lug nut lying around so I put it on the chip and it ran fine for as long as the computer ran.
Re:Ehh, it's been done before (Score:4, Interesting)
Another one was one of my own, that was near a window, and that side of the room got very cold in the winter. My systems always run 24/7 because this way the internal temperature stays somewhat consistant, avoiding chip creep and spreading solder joints. But then one day, when it was particularly cold out and so also very cold in that corner, I wanted to move another hard drive into it. At first it powered back up for about 30 seconds then shut down. Tried starting it several times but each time the running interval got shorter, until finally, it just wouldn't turn on at all. Replaced PSU and all was fine. Temperature change killed it.
A third one, this time the one in my gaming rig, developed its problem while in use. I was playing Oblivion or something intensive like that, and it was summer, very hot outside and in - my apartment is very poorly insulated as you may have guessed by now. The system started shutting down about every half hour, so after a couple instances of that I stopped playing, but later in the evening when things had cooled down, it was still doing it. Replaced the PSU and it ran fine after that. Temperature killed it.
Quite a few hardware failures I've encountered, CPUs, hard drives, video cards, whether my own or friends or work-related, I've been able to blame on temperature one way or another.
Re:The rule of thumb is.... (Score:2, Interesting)
I'm not completely certain on that explanation, but that's the general idea.
Re:hardhack? why? (Score:4, Interesting)
Hardhack is short for hardware hack.
Re:Ehh, it's been done before (Score:3, Interesting)
Not as dramatic but equally dumb: a friend had a small-form-factor Compaq Deskpro. Very tight little case. Shipped with a PIII/500 but he bought it used with no CPU. He decided to upgrade to a PIII/800. He bought one that was for a regular Deskpro and of course it didn't fit--so he used a Dremel to grind away almost half of the heatsink. Let's see... more-powerful chip, smaller heatsink, small case with limited airflow... I told him I didn't think it was a good idea, he said it wouldn't be a problem. He did the operation on Monday or Tuesday and it was dead by Friday.
And mine survived - The article's CPU might, too (Score:3, Interesting)
Still as it was unattended, the sever was left on the whole afternoon. I only realised it wasn't responding in the evening. The power was still on.
The heat of the processor evaporated the cooling liquid, and melted the plexi top of the CPU block.
And I still burned my finger when detaching the remaining copper block from the CPU even after a couple of minute after shutting down the power.
But even after all that cooking, once I replaced the cooling bloc and installed redundant pumps from a real brand (2x Lain DDC), the same CPU and motherboard started happily without complaining.
Now that's something that you won't be seing with more recent and fragile CPUs.
But given that VIA CPUs don't eat big amount of watts and don't generate high amounts of heat, it is still possible the their pico-ITX board will similarily survive heatsink-less cooking. The system will crash, but the CPU & ITX may be recoverable after letting cool down.
Re:Ummm, copying VW folks? (Score:1, Interesting)
Re:Ehh, it's been done before (Score:4, Interesting)
One morning I woke up and noticed my computer was off. I never turn my computer off. So I did what came natural, I turned the fucker back on. The powersupply blew the fuck up! I mean Boom! There was smoke billowing out the side and flame shooting out the fan port. The fucking fan in the power suppy was on fire! It was cool as fucking hell!
Fried everything in the damn computer but the CPU, memory, and graphics card. Harddrives, cdroms, tapedrive, ethernet/sound card, MB.. Gone. The only thing I can think of is it fried everything on the +5/+12V connection and since the surviving parts where 3V they lived. That is the best I can come up with.
Re:Ehh, it's been done before (Score:3, Interesting)
I regularly use that kill-a-watt meter on my home made pc's. I tend to build in the order of 10 pc's a year (just personal use; yes, I'm a member of hardware-anonymous but I stopped going to meetings.)
most minimal pc's (non gaming, like business 'web' pc's) tend to boot up at about 100w and lower down to 75w when the disk parks and when its in speedstep (etc) style mode.
I've not once seen any kind of low end pc get anything even close to 50w or less.
also note the via pc I measured did not have the most efficient PS. the pico style PS (which looks like a molex header with wires!) is said to be more efficient.
I believe I also had 2 drives on that box at the time. 2 notebook drives, that is.
for a 1ghz cpu (that does mythfrontend and mysql backend plus HDTV recording, and samba and nfs and
so lets say a regular pc is 75w and this pc is 45w. for an always-on pc, the cost savings DOES add up. not to mention I have slightly less a/c and cooling bills since my living room is not being heated up by a more monstrous cpu and fans.
Re:Done, accidently, before (Score:3, Interesting)
Sorry, very old assembly code joke that most will miss, but couldn't resist.
Re:Ehh, it's been done before (Score:4, Interesting)
My 3 favorite tools are made from paperclips.
The power supply tester
Creation: Unbend a paperclip, and then bend it into a big U shape.
Usage: When you are unsure whether a PSU works (a) disconnect it from anything it is connected to (b) insert one leg of the U into the hole in the 20/24 pin motherboard power plug for the green wire (c) insert the other leg into a hole for a black wire (d) plug the PSU into power and turn it on.
If the fans spin up, then the PSU at least partially works. At this point you can use a multimeter to verify the voltages of the different rails with no load.
The CD ejector
Creation: Straighten a thick-gauge (strong) paperclip, and then put a loop in one end that is big enough to put your index finger through, at least to the first knuckle (this helps with gripping it during use).
Usage: When you need to eject a CD from a powered-down computer (laptop OR desktop), push the paperclip into the emergency eject hole. On a laptop, this requires very little force, but on a standard (5.25") Desktop CD-ROM drive, this will take quite a bit of effort.
The multipurpose grabber
Creation: Straighten a regular paperclip, and put a loop on the end, as you did for the CD Ejector. On the other end, put a 90 degree bend, 2mm from the tip.
Usage: You can use this tool to remove or move jumpers (very handy for IDE hard drives), and to remove stuck floppy disks from floppy drives (use the R/W hole or 1.44MB hole as an anchor point).
Hope that was helpful to you.
Note: why use "creation" in the instructions? Well, I've had all the necessary components sitting in my drawer for years, and they stubbornly refuse to evolve into anything useful...