On 15 January 2014 15:23, Toby Corkindale <toby@dryft.net> wrote:
On 15 January 2014 15:08, Toby Corkindale <toby@dryft.net> wrote:
On 15 January 2014 09:59, Russell Coker <russell@coker.com.au> wrote:
A client is looking for a new embedded system. Their current systems have a 500MHz AMD Geode CPU (which provides a lot more power than they need). A randomly selected system has the following output of "free". So any of the bigger embedded systems (which seem to have 256M+) will do, but the small ones with 32M will be out of the question and even 64M probably won't do.
I can't speak for its quality or the stability of its Linux distributions, as I don't actually have it in my hands yet, but I've recently ordered one of these quad-core ARM boards and can let you know eventually. http://radxa.com/features/
Just following this up now, as I've had the Radxa Rock running for a little while. As I expected, I had to compile my own kernel and fiddle with the initramfs a bit before I could get it going with an ubuntu/linaro server filesystem and the kernel features I wanted.. but apart from that, it's pretty good as a mini server board. (The GPU features aren't accessible in Linux yet, sigh) It feels sooooo much faster than the 1ghz cubieboard or my 720mhz beaglebone, and those were already feeling like they were double the speed of the raspberry pi, which is just painfully slow. In a single-threaded app of mine, it takes 1.9 seconds to run on the Rock, and 21 seconds on the Raspberry. (Keep in mind that the Rock is a quad-core CPU, so for tasks that use them all, like video encoding, it will have even larger differences to single-core computers like the raspberry or beaglebone) -Toby