Custom PC, Issue 116

Custom PC, Issue 116This month’s Custom PC Magazine is a bumper issue for me: a massive in-depth Raspberry Pi feature is splashed across the cover, for which I provided all but the build-your-own-case section. As usual, the magazine also includes my regular interview column, this time talking to open hardware guru Andrew Back.

First, the Pi material. With the Raspberry Pi having had a phenomenally successful first year, and Custom PC having missed the chance to latch onto that with a cover splash for the launch review, it’s no surprise to see the magazine going all-out to attract Pi fans. Those who pick up the magazine for its Pi-related content are in for a treat, too.

First up is a head-to-head review covering the newly-released Raspberry Pi Model A and the redesigned Raspberry Pi Model B Revision 2. While some differences are obvious – the lack of a second USB port and Ethernet on the Model A, for example – others are less so, and the review hopefully answers the question of whether it’s worth paying the extra £12 to get the Model B over the tempting £18 asking price of the Model A.

The benchmarking continues with a look at how to overclock a Raspberry Pi without voiding your warranty, along with a few tips as to how to push it to ever-faster levels. Using a retail-model Raspberry Pi Model B Revision 2 equipped with a couple of cheap aluminium heatsinks, I was able to push the CPU from 700MHz to 1.1GHz and the GPU to 500MHz – and it made a serious difference in performance, as the benchmark results show.

Next, I walk newcomers to the project through installing the Raspbmc media server software and configuring it to stream HD YouTube content – something you’d think a £30 PC would struggle to do, but that’s certainly not the case. There’s also a look at the Minecraft: Pi Edition release, which provides a hackable and completely free version of Mojang’s popular block-’em-up game with which tinkers can fiddle around.

Finally, there’s a round-up of the four most popular operating systems for the Pi: Raspbian, the Debian-derived Linux distribution chosen as the ‘official’ OS by the Raspberry Pi Foundation; Raspbmc, the media-centric Linux distribution with integrated Xbmc support; RISCOS, by far the fastest OS for the Pi; and FreeBSD, for those who eschew Linux but still want a POSIX-compliant environment.

With the Pi work done, the interview. Andrew Back is one of the brains behind the Open Source Hardware User Group (OSHUG), and recently moved into my (relative) back-yard in Hebden Bridge. He’s a great guy, and always up for a chat – and his knowledge regarding open hardware, a still relatively unknown offshoot of the open source and free software movements, is second to none.

All this, and more by people who are not called Gareth Halfacree, can be yours if you just mosey on down to your local newsagents and pick up Custom PC Issue 116. Alternatively, stay indoors and get a digital copy via Zinio.

Linux User & Developer, Issue 104

This month sees just two pieces of mine in Linux User & Developer magazine: a review of the Efika MX Smarttop and a group test of popular filesystem-level encryption tools.

It’s always fun doing a hardware review for a change, and the Efika test was no exception. Based on an ARM architecture processor and shipping with a cut-down version of Ubuntu, any benchmarking tools need to be compiled from scratch in order to run.

The encryption group test was a departure from my usual fare: because the software on test has an actual, measurable performance impact on the host system, it’s possible to get an objective – rather than subjective – idea of which is ‘best.’

Building a custom benchmarking script, I created a small volume on a virtual host which was then encrypted using each of the filesystem encryption utilities. Files were then copied to and from the volume – with both sparse and dense files in small and large chunks chosen – with each transaction rated in terms of transfer speed and CPU load.

Between tests, the virtual machine was rolled back to an earlier snapshot – one of the major reasons I do this kind of testing within VirtualBox, rather than on a physical host – to ensure that file caching, fragmentation and the like couldn’t skew the results.

More information is available on the Linux User & Developer website.