46 comments

[ 311 ms ] story [ 212 ms ] thread
"Given how important chips are to modern society EVERYONE should understand and appreciate how they are made"

Everyone ? Maybe everyone in the field of chip making ? How does the knowledge of these definitions impacts the life of those not in this field (not being sarcastic, this is an actual question) ?

How can you fix your home electronics if you do not know all 200 chip designs??
Don't worry, most of the new "IoT smart devices" will probably be DRM-enabled, so you'll be a criminal if you try to fix them. Problem solved?
I'm using violence. Works when video starts flickering on our Samsung's TV.
I believe the article was satire.

I believe it was intended to mock developers evangelical quests to educate the public on programming languages/frameworks/methodologies/etc.

Sorry - there is no way this article is satire. Can you quote anything outright satirical other than the eagerness of the very first sentence? I reread it carefully in response to your comment but concluded there is nothing satirical about this article - on the contrary, it's a useful resource.
Agreed. Part of the problem is there's virtually no limit to the application of sentences like this. For example, "Given how important the Sun is for our survival, here's a list of 200 terms everyone should know about nuclear fusion."
Cool list(s), but I think the title is misleading. It should be something like "200-word chip design glossary", or "Chip design terminology: list of 200 definitions".

Now it sounds as if there are 200 actual chip definitions, whatever that would mean, since it doesn't say its (just) about words and jargon used in the field of chip design.

I confuse easily.

I believe it's a joke about programming language names/framework names/etc.
You're right - I though there where going to be circuit diagrams, or possibly pin-outs...
Common chip-design paradigms? Architecture patterns? I was expecting things like "TTL", "Discrete Logic", "Von Neumann", "Microcontroller" etc. etc. Interesting nonetheless.
I believe you're meant to click on the links to learn more. The links are more than just a definition - they're the Wikipedia article. Going through this list will end up amazingly useful for anyone who requires this level of glossary.
I assumed it was going to be snippets of VHDL, which is on my to-learn list.
looking at parts and datasheets to build a smps (switch mode power supply) and trying to figure out the networks that are rather complex from my point of view, I was expecting a list of 200 famous ICs like the lm317 or ne555 and when that seemed implausible, becuase there aren't that many iconic ones, I thought of insights into cmos design, which this hardly adresses and where 200 items would be a short list.
I'm not surprised they think everybody should understand how chips are made.

Using their documentation last year was a nightmare with errata in archived forum posts. I also found the community response quite hostile to simple outsider questions.

Then again, when we finally completed our testing of their chips we found the memory transfer overhead killed any benefit for us.

Out of interest: what was your intended purpose and what alternative did you decide on?
Some real time image convolutions with histeresis. We ended up switching to a Xeon Phi and abandoning the portable version.
It's a bit sloppy - the most egregious are:

Audio codec - compression is now common but only coding/decoding is required (hence the name).

GPIO - General Purpose Input Output.

Schmidt trigger - I have no idea what they wer trying to type but my cats do the same thing. A device with a Schmidt trigger input provides hysterysis for noise rejection?

I don't understand your sentiment, how are those egregious?

The referenced wiki article on audio codecs also talks about compression, in audio the two on very similar. GPIO is GPIO, but an engineer should know what they are and do. Schmitt triggers are really spelled like that, not with ..dt. Schmitt triggers are used for generating a stable block signal out of an analog signal, in order to decrease ripples and glitches

The copy is egregiously sloppy in places. The title is surely the most problematic, but the entry for Schmitt Trigger might be the worst:

    Schmitt Trigger: Comparitor circuit wityh
The list is decent, and intentionally superficial -- it's a list, not a textbook. The guy seems to know his material, but English might not be his first language, and I don't think proofreading happened.

I would guess that the list grew out of years of conversations and exposed gaps of shared knowledge. It's a good list of things that establish a common vocabulary in a specific industry.

Hm. What is the point? If you are in the field, you'll know most of these terms naturally. If you aren't, then you'll just forget them after passing the exam, or whatever.

Sounds like extra-thick satire to me.

If you want to get into the field then the Parallella board is good value, the Verilog and build scripts for the default configuration are on Github.
I kind of don't like them since they failed to deliver on their 64 core board.
Quite the contrast to the little piece of paper I have sitting on my shelve that reads

one and one zero

everything else is one

make a computer

;)

Nand?
Yes, the only haiku I ever wrote :P

I suppose "The Elements of Computing Systems" is an inspiring book.

If it was going to be real satire, they would have stuck at least one reference to a unique chip, say the Lays Wavy with Old Bay Spice. I'm going with it was a chip glossary that the author thought important.

Not the best title. I was also thinking it was the top 200 chips like the basic 7400 family, 6502, 8051, 555, etc.

ICs are one of the most important and impressive achievements of humankind imo. Taking a handful of Si atoms to form a transistor then carefully arranging a billion of them to make an x64 processor, just so Randall can watch a cat jumping into a box and make a webcomic of it.

Making useful things at those scales amazes me everytime.

I'm impressed too but also dismayed by how little the average person (or even the average CS person) understands about the low level stuff. I had a top notch CS undergrad degree, but I'm only truly able to grok some of the finer points 15 years out. And that's because I'm learning to use FPGAs as a hobby.

The Internet is truly marvelous in this regard. I'm using free tutorials, documentaries and videos to satisfy my curiosity and pretty much getting an EE education. Ignoring the car pictures, the Internet gives me much hope.

I'm in a similar boat. I have a CS background but only took one "Computer Organization" class which wasn't even that good but it gave me a starting point. I wish I'd at least minored in EE. What resources have you found especially useful? I'm using All About Circuits[1] to learn fundamentals of electricity in addition to whatever tutorials I read for particular projects.

[1]: http://www.allaboutcircuits.com/

Do you have any hardware you would recommend?
I usually purchase a lot of electronics hardware on a whim. With FPGAs, I'm forcing myself to learn how to use them before I spend the cash. There are quite a few boards out there though .. but this isn't a rec based on actual usage. The Papilio Duo seems like the one I'll get. My main concern is that this (and other hobbyist boards) use the Spartan 3 or 6. From what I can tell, it seems this has been retired by Xilinx. I'd prefer going with a Xilinx since the toolchain seems more popular with hobbyists.

Oh .. and for now, I have a cheap Cypress PSOC that I'm playing with. They were literally giving 'em away at Maker Faire :)

I didn't study CS, rather physics, and I find I know far more about "how it works" than most CS graduates I encounter and/or hire.

The reason is straightforward enough - I was a kid, and am now an adult, who cannot stop asking "why?", and habitually disassembled machines to understand them (and then usually hastily reassembled upon hearing the crunch of parental tyres on gravel). Having access to Brittanica as a kid helped too. I started making useful analogue circuits before I even knew ohm's law, as I could see and grok what components did when put together. Nothing beats reverse engineering as a learning experience, apart from perhaps forwards engineering.

Essentially, to have deep knowledge of something requires a desire to have that deep knowledge. Most are willing to stop at some level of abstraction and accept that things "just are", but others, myself included, need to understand how electrons work, why P-N junctions do what they do, what is this matter/energy thing anyway, or how about that spacetime thing that it sits within?

Turtles all the way down.

But why should the average person know how about "low level stuff"? A CS major should know the the level of gates and boolean logic but deeper than that why? Intel's cofounders weren't CS majors, one was physics and the other chemistry.
I agree. Unless the CS major is planning on specializing in embedded software, things like complexity classes and database normalization would be of more use. Even that is overkill, as in my experience leveraging and utilizing software libraries in the correct way often has already solved the issue at hand. Nowadays even more high level skills as that such as networking, caching, and distrusted computing are of more use when it comes to optimization.
For the same reasons they teach everyone history, math, languages, and the arts -- you should know on whose shoulders you stand.
Having spent my entire career in the chip industry designing and defining chips, I'm still amazed year after year.
Replace "everyone should know" with "someone may be curious to know", and the title will start to make sense.
And yet, somehow my grandmother managed to pull through without knowing about Amdahl’s law and wafer thinning
Lista are not that pedagogical. "Heres a French-English dictionary. start speaking French."
(comment deleted)
Original title is: "200 chip definitions everyone should know" OP had the good sense to change and strip it off the of self-importance and self-righteousness (although the writer probably meant to say 'everyone in IT business). Don't get me wrong. I'm not trying to be stupidly pedantic but no I do not think that social worker next door needs to know what ADC means.

IT-wise, yes younger generations seem to be in total loss with ICs and I believe the simple reason is that there are no jobs there anymore so there isn't a motivation to learn more about it. Virtual everything (as started off with JVM) has contributed too there. Not sure though that they are missing out that much.