Repairing the Itona VXL thin client

[Jim] was the happy recipient of 11 non working Itona VXL thin clients. The units he received were 800Mhz CPUs with 256 MB of Ram and 256MB of storage. None would power up. Upon internal inspection, he found a common theme. Leaky bulging capacitors in the power supplies. Since these came with custom 50W power supplies, he opted to simply replace the caps instead of replacing the supplies themselves. Now he has 11 fully functional units. There are great pictures and lots of info on his site, but what he doesn’t talk about is what he’s going to do with them.

Why don’t you pop on over to our Hacker Q&A and tell us what you would do with them.

Building a cluster of iPaq PCs

[Steven Pigeon] got his hands on ten iPaq computers that a friend acquired through an eBay auction. The older machines were in good condition but the march of technology had left them behind as casualties. He’s given them new life by assembling a cluster. The first order of business was testing the hardware to make sure it’s working. [Steven] used memtest86+ that comes along with the Ubuntu distribution of Linux to find one bad memory chip in the bunch (a revelation that took 10 hours to discover on the slow hardware). He assembled the unit above with MDF as a support structure¬†and threaded rod to hang the boards. He ended up with a beautiful module and his next step is to choose the operating system that will pull the whole thing together.

We find this build every bit as beautiful as the file cabinet cluster. It’ll be interesting to check back with him and see what kind of performance he can get out of it.

Distributed computing in JavaScript

mapreduce

We’ve heard about the idea of using browsers as distributed computing nodes for a couple years now. It’s only recently, with the race towards faster JavaScript engines in browsers like Chrome that this idea seems useful. [Antimatter15] did a proof of concept JavaScript implementation for reversing hashes. Plura Processing uses a Java applet to do distributed processing. Today, [Ilya Grigorik] posted an example using MapReduce in JavaScript. Google’s MapReduce is designed to support large dataset processing across computing clusters. It’s well suited for situations where computing nodes could go offline randomly (i.e. a browser navigates away from your site). He included a JavaScript snippet and a job server in Ruby. It will be interesting to see if someone comes up with a good use for this; you still need to convince people to keep your page open in the browser though. We’re just saying: try to act surprised when you realize Hack a Day is inexplicably making your processor spike…

[via Slashdot]

A Basic stamp supercomputer

basic_super

Hobby super computer building isn’t something you hear about every day. This project is even more peculiar due to the fact that it is a supercomputer built with BASIC Stamps. [humanoido] posted some great pictures and detailed info about his project. We’re not completely sure what definition of supercomputer he’s using, but he states that it beats out the others in 10 categories. Those categories are: smaller, lighter, portable, field operable, runs on batteries, has greatest number of input/output, has greatest number of sensors/variety, lowest power consumption, lowest unit cost, and easiest to program. Those sound a little more like features than supercomputing categories to us, but that doesn’t detract from the fact that this is one cool jumble of wires.

You may be wondering what it does. Well, so are we. From what he says, it talks in Chinese and English and has a plethora of other input and output devices. It also displays status of its internal communications. Catch a video after the break.

Continue reading “A Basic stamp supercomputer”

Another IKEA linux cluster

In a bit of serendipity, reader [Tim Molter] had decided on the IKEA Helmer cabinet for his new cluster right before seeing the previous IKEA cluster we covered. He and his coding partner recently completed building their own version of the IKEA Linux cluster. The cabinet was $30 and holds six headless boxes. Each board has a quad-core AMD processor for a total of 24 cores. They also feature 1GB of RAM and an 80GB laptop SATA drive. The latter was chosen because of space limitations in the case. [Tim] describes the Helmer cabinet as being almost perfect. The power supply lines up with the top edge of the drawer and the motherboard fits with a millimeter to spare. Power buttons were added to the front plus slots for airflow. It looks like a really clean installation and at $2550, incredibly cheap for the processing power.

IKEA Linux cluster


Building a render cluster doesn’t mean you have to spend a lot of money, even if you’re buying brand new hardware. [Janne] built this 6 unit cluster inside of a 6 drawer IKEA Helmer cabinet. He wanted the cluster to be low power and low cost. After finding a good price on 6 65nm Intel Core 2 CPUs, he found 6 cheap Gigabyte motherboards. The memory on each board was maxed at 8GB. With 24 2.4GHz cores consuming 400W, the power consumption and cost isn’t much more than a high end PC. Each board is running Fedora 8 and mounts an NFS share. Dr Queue is used to manage the render farm’s processes. [Janne] says jobs that previously took all night now only require about 10-12 minutes. The estimated capacity is 186Gflops, but plans are already in motion for a12Tflop version.

His site also has plans for an underwater camera housing like our recent post. If you want to see more IKEA abuse, check out IKEA Hacker, even if it’s not very technical.

[via Hackzine]

[UPDATE: yep, we duped ourselves]

24 core Ikea cluster


[Janne] does freelance animation and wanted something with a bit more CPU to get his rendering jobs done. He picked up an Ikea ‘Helmer’ cabinet and refitted it to hold six Intel quad cores, six Gigabyte motherboards with 8GB of ram each and six 400 watt power supplies. He seems happy with it – I think it just needs some custom power wiring and an integrated Gig-E switch to achieve perfection. What? I’m not jealous at all.