It’s no secret that I’ve backed the Pine64. I’ve waxed geek about that already. But why did I choose this platform for the clustering I have in mind?

I had been looking for the next personal project’s hardware for about a month after taking a few months to decide what I actually wanted to do (more on that some other time). I reached out to my network on Facebook at the time the Pi Zeros were coming out. Many of us were excited about this for various reasons. For me, it was a matter of cheap hardware for a cluster. I still wasn’t that excited. I just saw more hardware for my dollar.

Jon “Maddog” Hall, who I met back in 2005 in Boston at LinuxWorld, had other thoughts and interceded- I’d tagged him because I knew he was doing things in the Pi-sphere. He wrote a comment initially, and I’m guessing he had to do it one too many times and wrote an article: “The Run on the Raspberry Pi is Temporary. Don’t Panic.” The part that caught my attention:

…On the other hand, you can purchase an eight-core, 64-bit ARM system with two GB of RAM and wireless 802.11n built-in, running at 1.2 GHz for $99 [1]. The efficiency of OpenMP or OpenCL to control your parallel threads on a shared-memory model for your application instead of OpenMPI across networking interfaces will be breathtaking (this is not a slight on OpenMPI; every tool has its place)…

64 bit. That certainly would make my nodes more powerful singly and in parallel. I started looking at the 64 bit options in the Pi-Sphere and found nothing that really tugged at me. The market remains a bit of a “well here it is, as is” space and time spent researching allowed other new things to pop up. It was taking longer to research than to actually get started, and that paralysis was something I simply not accept from myself (I hate it enough in professional settings). And then I found the Pine64. It smells of potential, so I backed it with the 2 Gigabyte of RAM option so that I could get 10 of them. And in April I should have them.

It seems that when we go out to look for things, we look for familiar things – that’s probably why McDonalds in many countries serves so many tourists in other countries – the (misplaced?) comfort of something we all know. It’s what keeps people using the same technologies.

That’s pretty boring. I never have done boring well and I do have references.

So that’s why I picked the Pine64, and so far – despite my lack of hardware to this point – I’m using the time to plan things out for when they get here. And as it happens, through their updates, they’ve been doing some of the legwork for me by their partnerships, particularly the one with PicoCluster LLC. That one takes some will maybe take some of the administrative aspects of the clustering away.

So I await with very minty breath. While I’m waiting, maybe you should consider backing the Pine64 too.