Shop OBEX P1 Docs P2 Docs Learn Events
DefCon — Parallax Forums

DefCon

JonnyMacJonnyMac Posts: 9,194
edited 2012-08-01 17:24 in Propeller 1
Ryan Clark, another former Parallax employee, hired me to play a spy character at DefCon as part of his cryptography game. While not providing clues (valid and bogus) to participants, I networked with lots of hardcore hackers (like Joe Grand) and even taught three introductory Propeller classes. That was the intent, anyway -- these were really advanced folks and they were happy to jump in and explore assembly (to the degree I'm proficient). I think the classes were a success; they were scheduled for two hours and all went three plus.

The take-away from this event is that it is important for Parallax to create a more professional tool chain with C compiler. Most loved the Propeller and said so; they weren't wild about the tools (used Propeller Tool and BST for demos). Many that asked about tools expressed an interest in using GCC which seems to be good news for the team working on that.

Sadly, I had several people come up to me and say, "Until now, I've never heard of the Propeller." Yes, it seems strange, but I think it offers a marketing opportunity for Parallax. We're all very good at preaching to our own choirs, it may be time to visit other "houses of worship."

Related to that last comment, I had a few people come up to me when I was in the Hardware Hacking Village working with a QuckStart board and ask, "Is that a new Arduino board?" Oi... As DefCon is a rough-and-tumble bunch I went to the vendor area where I could get an instant custom shirt. Mine says:

MY PROPELLER KICKS
YOUR ARDUINO'S @SS!

I figured if nothing else, it would start a conversation where I could explain facts and dispel myths (like one young man running around claiming that the Propeller only time-slices, it doesn't use multiple processors).

I also spoke with the owner of a publishing company about writing a Propeller book (using Spin, for beginners); we'll be discussing that further in the coming days.

Comments

  • SapiehaSapieha Posts: 2,964
    edited 2011-08-08 10:52
    Hi Jonny.
    The take-away from this event is that it is important for Parallax to create a more professional tool chain with C compiler. Most loved the Propeller and said so; they weren't wild about the tools (used Propeller Tool and BST for demos). Many that asked about tools expressed an interest in using GCC which seems to be good news for the team working on that.

    I to will have C but for me it is NOT what it will be LCC, GCC else other - Important functional.

    BUT if Team that build it CAN'T find way to program propeller that simple in as it is in spin in Parallel use of COG's --- Instead convert Propeller to Single COG ---- Very Bad!

    And As I see Them still NOT presented any PROGRAMING MODEL to that.
  • JonnyMacJonnyMac Posts: 9,194
    edited 2011-08-08 11:02
    I spoke with a Brazilian professor who sat in on my class and LOVED the Propeller architecture, though he really wants C (he taught a class on C for the PIC during DefCon). In our discussions he mentioned something about SDCC (small device C compiler). I'm not a C gearhead so I don't care very much, but, perhaps,those working on C for the Propeller will.

    BTW... that professor says he will be in contact with Parallax as he does want to integrate the Propeller into his curriculum.
    BUT if Team that build it CAN'T find way to program propeller that simple in as it is in spin in Parallel use of COG's --- Instead convert Propeller to Single COG ---- Very Bad!

    Agreed; any language developed for the Propeller has to support multi-processing using multiple cogs.
  • potatoheadpotatohead Posts: 10,261
    edited 2011-08-08 11:34
    Love the shirt!

    Yeah, perfect for that crowd.

    IMHO, Spin needs more promotion. I love it, it's lean, it's flexible, etc...

    Thanks for sharing your impressions.
  • rod1963rod1963 Posts: 752
    edited 2011-08-08 19:23
    Looks like Parallax is on the right track porting GCC/Eclipse to the P1 and P2.

    People come to expect it nowadays since it's available for almost every new micro released.
  • jazzedjazzed Posts: 11,803
    edited 2011-08-08 20:27
    JonnyMac wrote: »
    Agreed; any language developed for the Propeller has to support multi-processing using multiple cogs.

    That's part of the current plan by threading like posix pthreads and via PASM.
  • cavehamstercavehamster Posts: 5
    edited 2012-07-31 22:37
    Speaking as one of those folks from defcon, and this being my first real experience with the propeller chip, here's my thoughts:

    1) Not a huge fan of spin or any language where whitespace matters. I want to be able to hack out code in any random IDE and not worry about it messing up the formatting.
    2) I like gcc; it runs on everything. My current process of using Atmel AVR chips is to use the Xcode IDE under OSX to develop, using make and gcc to compile and program my part
    3) The propeller chip seems like a fun little device. I've already planned to swap to it for a little project I have coming up. Wish it did USB natively though.

    Parallax and Ryan did a great job with the badges this year, they were very accessible to even folks who had never handled a soldering iron. I know I met a lot of new friends over hacking away at the badge challenge.
  • pedwardpedward Posts: 1,642
    edited 2012-07-31 23:03
    I'm getting the sense that even the Propeller faithful don't have a really good idea of what Parallax and crew have been up to.

    GCC for the Propeller is here, now, and it is really slick. The GCC runtime for the Propeller supports debug serial I/O, simple serial, high speed serial, SD FATfs support, and POSIX pthreads.

    Steve wrote a really fantastic IDE for PropGCC that makes using the Prop 100x simpler than any other vendor's toolchain. It's simplicity is on par with the Arduino development environment, but is written in C++ and available natively for Windows, Linux, and Mac.

    It's a shame that this wasn't well represented at Defcon.
  • David BetzDavid Betz Posts: 14,516
    edited 2012-08-01 04:50
    Sapieha wrote: »
    BUT if Team that build it CAN'T find way to program propeller that simple in as it is in spin in Parallel use of COG's --- Instead convert Propeller to Single COG ---- Very Bad!

    And As I see Them still NOT presented any PROGRAMING MODEL to that.
    In fact, this is already supported by pthreads as well as by more Spin-like cognew and cogstart runtime functions. What is not currently supported is running code out of external memory on more than one COG at a time. You can already run LMM code on as many COGs as you like with as many threads on each COG as you like. While we will support running XMM code on multiple COGs at some point (maybe soon), I doubt that will be used very often since performance penalties associated with putting code in external memory are likely to be multiplied by trying to share the same external memory interface among multiple COGs.
  • TymkrsTymkrs Posts: 539
    edited 2012-08-01 17:24
    JonnyMac wrote: »
    I also spoke with the owner of a publishing company about writing a Propeller book (using Spin, for beginners); we'll be discussing that further in the coming days.

    If you guys need any help or need any reference to a beginner learning Spin, hopefully you've checked out http://firstspin.tv
Sign In or Register to comment.