An interesting observation
Skogsgurra
Posts: 231
It is Saturday.
We were discussing some interesting topics. I happened to mention that I am in love with a new chip. People stared at me - not believing they heard me right. I explained. Got a piece of paper and drew a sketch explaining the Propeller architecture...
I than dawned upon me: Hey! I am drawing this complete architecture just like that - from the top of my head. Never done that before. What's going on here?
Simple. With such a straightforward and uncluttered design, there's not much to memorize. It just comes naturally. That is also a quality that not many other chips have.
Had to share...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
We were discussing some interesting topics. I happened to mention that I am in love with a new chip. People stared at me - not believing they heard me right. I explained. Got a piece of paper and drew a sketch explaining the Propeller architecture...
I than dawned upon me: Hey! I am drawing this complete architecture just like that - from the top of my head. Never done that before. What's going on here?
Simple. With such a straightforward and uncluttered design, there's not much to memorize. It just comes naturally. That is also a quality that not many other chips have.
Had to share...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Comments
A window is almost as good for geek status, UV pen optional.
But seriously it does have a certain elegance.
Graham
Never thought about it, but you are right! I've never diagrammed a chip like that for discussion before. Networks, data structures, etc... all see frequent diagrams, but never CPU's.
Where did your conversation go, after that? In mine, I found after laying out the chip, the discussion settled on multi-processing and parallism. Differentiating the Prop, from other things like multi-core, NUMA, those networked clusters I can't seem to remember the word for, etc... proved difficult.
The idea of dedicating COG's to tasks, to form what I like to call 'sardware' appeals quickly. People grok that. IMHO, that's probably the easiest selling point to get across. (Which is why I really like, "The Power of 8" for the new slogan.)
Determinism proved much more challenging. The average person, interested in this stuff, will almost always ask, why not just have one fast CPU, then these COG's? Or they ask, why have so much on the chip? Of course, the reality is there isn't that much on chip, it's just very easily re purposed!
Just wondering where that thread went for you Skogsgurra. What did they grok right away, and what proved hard?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Propeller Wiki: Share the coolness!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I stopped caring what "normal" people think of me or of people like me. I like chemistry, electronics, history, languages.. and pr0n. If they don't, I can't help them
Post Edited (Ale) : 9/15/2007 8:24:57 PM GMT
Discussions go wrong when you do not distinguish between:
- What current technology can do ("I shall now show you a movie where people repair the ISS space station")
- What you can buy for resonable money ("Look at this 3 Euro MP3 player with display and storage for 1000000 songs")
- What you have done yourself ("Here is my new 'Turtle Robot': It will obey to the commands: 'Come', 'Hide', and 'Play Music")
Maybe a Propellor is involved in each of these scenarios, or not - I have never managed to explain the significance of this design layer.