Shop OBEX P1 Docs P2 Docs Learn Events
40 Core Processor with FORTH IDE — Parallax Forums

40 Core Processor with FORTH IDE

DroneDrone Posts: 433
edited 2008-10-06 11:43 in General Discussion
Has anyone looked at this 40 core processor with a FORTH IDE? How does it compare with Propeller I and what we know about II.
Via Elektor Newsletter No.186: said...

IntellaSys has announced its SEAforth 40C18, a 40-core multicore processor designed for embedded control in wireless, portable, and distributed data processing applications. The SEAforth 40C18 can execute 80% of its VentureForth instructions in 1.38 ns with a power consumption of 7 mW or less per CPU.

Web Site:

www.intellasys.net/

Datasheet:

www.intellasys.net/templates/trial/content/S40C18_DataSheet.pdf

Comments

  • hippyhippy Posts: 1,981
    edited 2008-10-03 14:23
    I haven't really studied it in detail but things which strike me as fundamentally different to the Propeller -

    1) Cores are in a fixed matrix only able to communicate with non-diagonal adjacent cores.

    2) Cores on the periphery have dedicated intended I/O functions. Central cores have no direct I/O capability.

    3) The Propeller is simply 8 MCU's with shared memory, this is a multi-core configuration which needs designs which allocate specific tasks to specific cores with specific pathways creating between and through other cores.

    4) It is intended to be programmed in Forth. Nothing wrong in that per se but many programmers do have a difficulty in using Forth or a resistance to it as a first-choice programming language.
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2008-10-03 14:41
    This thread is not Propeller related and is being moved to the Sandbox Forum.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Engineering
  • davejamesdavejames Posts: 4,047
    edited 2008-10-03 16:02
    FORTH?!?!?· Wow!· That's a "blast from the past".· I haven't heard hide nor hair of that language since the mid-80s.· Is it really still in use?

    Interesting language, but not sure if there's any advantage to its "word-extension" vs. a "function call library" approach.

    DJ
    ·
  • LeonLeon Posts: 7,620
    edited 2008-10-03 17:51
    It's still popular with some people. A Forth for the Propeller was being developed.

    I/O drivers for SUN workstation peripheral boards were written in Forth, and there was a Forth system in the machine. We had some SUNs where I once worked and I found how to get into the Forth system on them. I think a key had to be held down while they booted.

    I have a feeling that the Propeller might be more popular if it actually had an architecture more suited to Forth, and used Forth instead of Spin. Speed would approach that of assembly language. The designers probably considered it at one time.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle

    Post Edited (Leon) : 10/3/2008 5:59:18 PM GMT
  • davejamesdavejames Posts: 4,047
    edited 2008-10-03 19:14
    Well, I'll be!

    http://www.forth.com/resources/appNotes/index.html

    DJ
  • Carl JacobsCarl Jacobs Posts: 41
    edited 2008-10-03 22:15
    A Forth compiler is currently available for the Propeller. It takes forth source code, and compiles it to an object which can be linked with any other objects written in Spin or PASM for the Propeller chip.

    More details at:·http://forums.parallax.com/showthread.php?p=741699


    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Carl Jacobs


    JDForth - Forth to Spin Compiler http://www.jacobsdesign.com.au/software/jdforth/jdforth.php
    Includes: FAT16 support for SD cards. Bit-bash Serial at 2M baud. 32-bit floating point maths.·Fib(28) in 0.86 seconds. ~3x faster than spin, ~40% larger than spin.
  • pharseidpharseid Posts: 192
    edited 2008-10-04 02:03
    · The Seaforth chip came up earlier and although I like forth, the amount of memory per processor is measured in bytes, if I remember correctly. Like 64 bytes per processor or something. That seems awfully limiting. Especially when not all processors have direct access to I/O pins, as has been mentioned. (Okay I just checked, it's actually 64 18-bit words, but the same principle applies).

    ·-phar
  • LeonLeon Posts: 7,620
    edited 2008-10-04 12:46
    I can't really see the point of that SEAforth chip. I'd be a bit more impressed if they were actually selling the things. Chuck Moore keeps designing weird chips but I've never heard of anyone actually using them!

    XMOS is aimed at the same markets and their chips are actually being designed into products. Cambridge University is using them for student projects.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle

    Post Edited (Leon) : 10/4/2008 12:55:56 PM GMT
  • pharseidpharseid Posts: 192
    edited 2008-10-06 05:30
    · Intersil-Harris bought the rights to his Novix chip (the first Forth chip Chuck developed in the 80's). I liked his ShBoom chip, another simple design which was tuned to exploit the burst access mode of DRAM's. I suggested to his partner that it would have really been interesting if they implemented it on the same die as a DRAM and he told me that they had a deal to do that with one of the DRAM manufacturers but it fell through. I think that might have found it's own niche in the MCU market, a cheap fast chip with lots of RAM.

    · I know one of the selling points for chips like the XMOS chip are as replacements for FPGA's for people who don't want to write VHDL. I suppose the Seaforth chip would be the replacement for a CPLD.

    -phar
  • LeonLeon Posts: 7,620
    edited 2008-10-06 11:43
    There is a lot of discussion of the SEAForth chip on the comp.lang.forth NG, and someone I know slightly has one of their kits. When I get my XMOS kit I'll do some benchmarking, and someone on the NG can run the same benchmarks on the SEAForth chip.

    Leon

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Amateur radio callsign: G1HSM
    Suzuki SV1000S motorcycle
  • LeonLeon Posts: 7,620
    edited 2019-09-09 14:24
    I have come across an IntellaSys kit I'd completely forgotten about! The USB I/F is recognised by Win 10 - the module comes up in the device list. IntellaSys is no longer with us, www.intellasys.net is now "Intelligent aids for daily living".
  • PublisonPublison Posts: 12,366
    edited 2019-09-09 14:38
    Leon. Good to see you back!
    I was able to look at some old websites from IntellaSys using the Wayback Machine.
  • VonSzarvasVonSzarvas Posts: 3,450
    Yes. Good to read your update, Leon.

    Welcome back, and hope you are feeling well.
  • SaucySolitonSaucySoliton Posts: 521

    A similar chip with 144 cores is available from GreenArrays. In fact, some of the IntellaSys engineers are there now.
    greenarraychips.com/home/about/bios.html
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2019-09-10 13:49
    SaucySoliton wrote: »
    A similar chip with 144 cores is available from GreenArrays. In fact, some of the IntellaSys engineers are there now.
    greenarraychips.com/home/about/bios.html

    I've always been fascinated by these chips but can't see how I can use them effectively. So I adopt the waiting attitude to see what the manufacturer comes up with. After all these years there is still nothing in the appnotes other than PS/2 keyboard interface and 10baseT NIC etc. Really!? Not even USB keyboard or 100baseT?
    That does not impel me to purchase any chips in lots of 10 or a dev board for $500.

    Chuck should have been able to come up with some killer app for these chips, but even he hasn't.
  • David BetzDavid Betz Posts: 14,516
    SaucySoliton wrote: »
    A similar chip with 144 cores is available from GreenArrays. In fact, some of the IntellaSys engineers are there now.
    greenarraychips.com/home/about/bios.html
    This chip looks pretty interesting but the eval board is expensive.

Sign In or Register to comment.