pointless precision

posted by tom / April 19, 2006 /

Let's nip this in the bud right now: just because an electronic device has two different embedded circuit packages in it does not make it "dual core". I'm going to find it pretty irritating (admittedly, for no good reason) if "dual core!" stickers start getting slapped on every new consumer gadget. In this case it's a particularly bad example: having a digital/analog converter outside the general-purpose processor is completely normal and unremarkable. If you were building an mp3 player out of off-the-shelf components (that weren't specifically designed as integrated solutions for building mp3 players), you'd end up with a separate DAC quite naturally — it's not exactly a premium feature. And you'd really be a dope to call such a solution "dual core".

But if you insist on it, you might as well call the iPod sextuple core. Check it out: it's got a genuinely dual core ARM processor, a DAC, a firewire controller, an LCD controller and (not listed on that page) I believe the video on newer models is handled on a separate IC. What does all this mean? Nothing, other than that — surprise! — electronic devices aren't magic boxes. They have pieces inside of them. Unless you know what they are, counting them won't tell you very much about which gizmo to buy. I suggest sticking with the traditional "shiniest" heuristic.

Comments

Ooh, a post right at the outer reaches of my nerdiness comprehension---I dig.

Tom, you have my word: I'll help wage the war against ignorance ("the long war?") here on the front lines of consumer-electronics-ville.

Posted by: Stanley on April 19, 2006 04:29 PM

There's always the "most expensive" approach.

Posted by: Doug on April 19, 2006 04:47 PM

Because I'm sure you're not reading that thread anymore...

How would a button on the comments page that triggers some javascript that requests only new comments from the server and then dynamically adds them results in increased server load? Because it's not using the cache of the html of the comments page in that case?

I suppose, maybe. It would help on a wordpress site, though, unless wordpress does caching and I'm not aware of it.

Posted by: pdf23ds on April 19, 2006 05:13 PM

Sorry I didn't check back in on that thread, pdf. But yes, because of the caching concern you outline. Under the JS scheme the server would have to figure out which comments count as new, which introduces a much larger load than just shoving out a pre-computed file representing the current state of all the comments (since the current bottleneck is processor time, not bandwidth).

Posted by: tom on April 19, 2006 05:45 PM

Reminds me of the TurboGrafx-16 purporting to be a 16-bit system when it was actually an 8-bit CPU with a 16-bit GPU.

Posted by: J. Miller on April 19, 2006 06:03 PM

Oh, Turbo Grafx 16, those fucks! Thanks for breaking it down, I totally get that.

Posted by: Kriston on April 19, 2006 07:59 PM

Post A Comment

Name


Email Address


URL


Comments


Remember info?



Google Analytics