Linked by Thom Holwerda on Sat 9th Nov 2013 18:44 UTC
Games

The recent news, pictures, and reports about the prototype Steam Machine got me thinking. Since the SteamOS platform is a simple x86 computer, without the kind of restrictions that regular consoles come with, you can simply build the prototype Steam Machine today. However, the big issue is that Valve has done some magic to make sure that the hefty processor and videocard are properly cooled in the tiny prototype enclosure.

For years now, I've been looking for a way to build such a powerful PC in such a tiny package. The problem is that building such a small, powerful PC yourself is not easy - especially not for someone like me, who doesn't have the time to keep up with the honestly irresponsibly large amount of options available in the processor, videocard, cooler, and case markets. It's a mumble-jumble of version numbers, and in the case of video cards, cooling designs, card lengths, and god knows what else.

So, I have a simple question. Say I want to build a small, powerful gaming PC like the Steam Machine prototype, using off-the-shelf parts, for a reasonable price (I would say EUR 600-800). It needs to be properly cooled and as silent as possible, and it needs to be a small console form factor - so a small, horizontal case. Building a powerful, cool gaming PC in a tower is easy. Building it small and console-like, however, is not.

So, if you were to build something like that, which components would you pick? I might - no guarantees! - take up the advice given here and actually build it, if I can justify the spending. Even if I don't - it seems like a nice exercise for the PC builders among us. The laptop, smartphone, and tablet explosion has pushed custom PC building to the sidelines, but I still think it's an incredibly fun and satisfying activity - and if you're good enough, it is, most certainly, an art.

Permalink for comment 576743
To read all comments associated with this story, please click here.
RE[3]: Comment by Luminair
by _txf_ on Tue 12th Nov 2013 23:07 UTC in reply to "RE[2]: Comment by Luminair"
_txf_
Member since:
2008-03-17

It appears you're correct.

Also as I recall the last time I heard that single channel memory limitations were related to a bobcat based apu (which only supports a single memory channel).

I was always under the impression that having two channels would increase performance, after all one could in theory hide latency by schedule reads and writes to each channel independently...

Reply Parent Score: 2