Forum Rule: Always post complete source code & details to reproduce any issue!
Results 1 to 4 of 4

Thread: Teensy 4 is really really fast ! :-)

  1. #1
    Member
    Join Date
    Oct 2016
    Location
    Paris
    Posts
    32

    Teensy 4 is really really fast ! :-)

    Hi,

    I am currently designing a Rubik's cube solving robot (not very original I know :-)). Anyway, I built a robot using a Pi-zero W but now I want something without the overhead of a full linux distro so this should be a perfect job for a Teensy 4.1... Therefore I ported kociemba's (beautiful) algorithm to Teensy 4.1.

    The original code is here.

    It turns out that the results are really impressive ! Solving a cube (in at most 24 moves) is about 25% faster on average on the Teensy 4.1 than it is on the Pi Zero with the same C code! Admittedly, I did try to optimize memory access for the moves cache tables on the teensy and not on the pi-zero and I am also running the teensy overclocked to 816mhz but it is still really impressive! That MCU is a beast :-)

    If anyone is interested, the port is available on github . The computation requires 4.3MB of flash memory to hold the pre-computed tables and, optionally, about 730Kb of RAM to improve cache access.

  2. #2
    Senior Member
    Join Date
    Jul 2020
    Posts
    174
    Interesting. How come you put the buffer in DMAMEM?

  3. #3
    Senior Member+ defragster's Avatar
    Join Date
    Feb 2015
    Posts
    13,495
    DMAMEM is just the other half of the 1024KB of RAM on the T_4.x. It is non contiguous so to make use of all RAM requires buffer in both free faster DTCM and the DMAMEM. In this case RAM between the two is 730KB.

  4. #4
    Banned
    Join Date
    Oct 2020
    Posts
    6
    very cool)))

Posting Permissions

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