Forum Rule: Always post complete source code & details to reproduce any issue!
Results 1 to 14 of 14

Thread: Google imxrt1176 coming soon

  1. #1
    Senior Member
    Join Date
    Feb 2020
    Posts
    139

    Google imxrt1176 coming soon


  2. #2
    Senior Member
    Join Date
    Apr 2014
    Location
    -
    Posts
    9,741
    Quote Originally Posted by TeensyWolf View Post
    Thanks for the link!

    The board looks very interesting. The CPU seems to be shielded(?!) - a good thing. EDIT: No, not shielded - looked at the wrong picture Also that there will apparently be an addon for WiFi / blutooth. I wonder if there will be Arduino for that? Too bad there is no detailed information (pinout etc) yet.
    I am very curious about the final price. The predecessor board for Linux with Cortex-A35 and GPU is sold here for a little over 100 EUR (Pollin). A bit much, the new board would not be worth that much to me. Especially since I don't really know what it could be used for - except AI (for which I can only think of "boring" purposes - what's the point if a face is recognized...). I don't think I'll ever need it ) But maybe there are other useful applications.
    I hope it will have a ready to use rtos.

  3. #3
    Senior Member
    Join Date
    Apr 2014
    Location
    -
    Posts
    9,741
    What I would like is to teach a robot to walk.
    But it doesn't increase the confidence in tensorflow when I see how badly the website is translated into German.... (I hope it was not translated by Tensorflow ;-)


    • Programmierkenntnisse: Gebäude ML - Modelle geht es um viel mehr als nur das Wissen ML Konzepte-es erfordert , um die Codierung der Datenverwaltung, Parameter - Tuning zu tun, und die Ergebnisse benötigt , um Test - Parsing und optimieren Sie Ihr Modell.
    • Mathematik und Statistik: ML ein Mathe schwere Disziplin, also , wenn Sie planen , ML - Modelle zu ändern oder von Grund auf neu, Vertrautheit mit den zugrunde liegenden mathematische Konzepte zu bauen ist von entscheidender Bedeutung für den Prozess.
    • ML - Theorie: die Grundlagen der ML - Theorie weiß , werden Sie ein Fundament aufbauen, und helfen Sie beheben , wenn etwas schief geht.
    • Erstellen Sie Ihre eigenen Projekte: Erste praktische Erfahrung mit ML der beste Weg ist , Ihr Wissen auf die Probe zu stellen, also nicht in früh mit einem einfach zu tauchen Angst, colab oder Tutorial etwas Übung zu bekommen.
    That's a pretty bumpy translation.
    Last edited by Frank B; 01-26-2022 at 01:45 PM.

  4. #4
    I was just about to post that info, but it looks like TeensyWolf was faster :-
    Product page for imxrt1176-based Dev Board Micro is now available: https://coral.ai/products/dev-board-micro/

    I don't know how easy would be to program the Coral Edge TPU on this board, but they advertise this accelerator is able to run 4TOPS (trillion operations per second). That is something like 1000x more processing power than ARM M7 processor at 1GHz.
    Imagine the algorithms that suddenly will become feasible in real time for this beast

  5. #5
    Senior Member
    Join Date
    Apr 2014
    Location
    -
    Posts
    9,741
    Quote Originally Posted by tomas View Post
    Imagine the algorithms that suddenly will become feasible in real time for this beast
    I would. I can imagine none that is not AI.
    So, which?

  6. #6
    Senior Member
    Join Date
    Jul 2014
    Posts
    3,473
    <Removed>

    Edit: Thanks Frank: it IS 64 MB RAM
    Last edited by WMXZ; 01-29-2022 at 07:10 PM.

  7. #7
    Senior Member
    Join Date
    Apr 2014
    Location
    -
    Posts
    9,741
    Not 64M?
    However I'm pretty sure T5 will have 0(zero) extra RAM, and without a RTOS (which seems to be fact) and two cores i don't expect a smooth running system. It may end in a way that you can use only one of the cores, with the slower wasting cycles in a unmaintainable state machine doing blocking spi, i2c etc and the faster has to wait for the slower..

  8. #8
    Quote Originally Posted by Frank B View Post
    I would. I can imagine none that is not AI.
    So, which?
    Everything that involves convolution is implementable on TPUs.
    Other algorithms too, check out some science on the subject:
    https://arxiv.org/abs/2107.05473

  9. #9
    Senior Member
    Join Date
    Apr 2014
    Location
    -
    Posts
    9,741
    Quote Originally Posted by tomas View Post
    Everything that involves convolution is implementable on TPUs.
    Other algorithms too, check out some science on the subject:
    https://arxiv.org/abs/2107.05473
    Thank you. This paper is interesting!

  10. #10
    Senior Member
    Join Date
    Sep 2021
    Posts
    170
    The gogle coral dev board "micro" with RT1176 is still not available.. still, the "pre-order" button is there.
    It has an "rtos" and would be interesting for that reason alone - without considering the camera and the TPU. Unfortunately, it doesn't seem to become a good distraction during the xmas

  11. #11
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,406
    Yup, supply chain problems are tough for everyone this year, even Google.

    The fact that PJRC has shipped any Teensy 4.0 & 4.1 in 2022 is a result of an pretty incredible amount of mostly invisible behind-the-scenes work, including several minor PCB revs to use alternate parts. Even then, the net result is meeting only about half of the total demand where the products are briefly in stock, then unavailable again, over and over...

  12. #12
    Senior Member
    Join Date
    Jan 2014
    Posts
    111
    Is Coral part of Alphabet/Google? anyway, here is their page for it. $80.
    https://coral.ai/products/dev-board-micro/

  13. #13
    Senior Member Blackaddr's Avatar
    Join Date
    Mar 2017
    Location
    Canada
    Posts
    344
    Quote Originally Posted by PaulStoffregen View Post
    Yup, supply chain problems are tough for everyone this year, even Google.

    The fact that PJRC has shipped any Teensy 4.0 & 4.1 in 2022 is a result of an pretty incredible amount of mostly invisible behind-the-scenes work, including several minor PCB revs to use alternate parts. Even then, the net result is meeting only about half of the total demand where the products are briefly in stock, then unavailable again, over and over...
    I can sympathize. I've been going nuts with my PCBA orders for the past year. Every time I submit another PO, I get a list of unavailable parts and start scouring datasheets trying to finding substitute parts that work (and are in stock!), then updating other parts for things like input/output capacitors, etc. It's exhausting.

  14. #14
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,406
    Just in the last couple days I've spent way more time than I wanted on voltage regulators.

    Even made this test on my workbench, since some of them have less than reliable info...

    Click image for larger version. 

Name:	vreg.jpg 
Views:	67 
Size:	173.2 KB 
ID:	29793

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •