defragster
Senior Member+
Manitou's perf post is a good one - it is nice to see how the MCU's stack up. Would be interesting to see an olde schoole 486 or Pentium-90 reference point. Interesting to see the T_3.1 boosted perf over the T_3.0 so much, and how the T_LC is really in a (ZERO) class between T_3.1 and UNO. Also how the K64 is closer to the T_3.2 than the K66 - perf-wise.
I also backed Kris/onehorse on his KS to support him - but not yet seen a link to his KS posted - Paul's was the first note AFAIK. The link posted above was in answer to a question about Arduino on STM? I've seen it done before where non Teensy users get told where to go - that isn't SPAM.
I ordered a T_3.6 MPU9250 board from Tindie just to see if I could figure out how Kris gets the T_3.6 to run as slow as the low power STM with near the speed of a T_3.2.
Between WFE and yield() in this thread and the ZILCH multitask thread that uses the same or fewer cycles in task switching by replacing the default that eats those cycles in yield() scanning all Serial# ports for serialEvent() calls I've learned some things about Teensy. This thread seems to have had a reason to start - where it went from not being ignored has some value in Teensy usage and perhaps areas to revisit for improved Teensy use cases.
<edit>:Just searched LadyBug - found one each in this thread and the MPU9250 thread - only link is to OSH PCB.
<edit2>: BTW : I also abused the kinetis_hsrun_disable() with my WFE sample and current dropped lower - but then USB comms were spotty as my simple example did it - mostly sleeping at low speed: the only waking interrupt I had was the systick, and time got 2::1 dilated going 240 to 120 MHz. I wondered if duff's Snooze speed mode change dropped the hsrun (by default or with code) - but didn't see an answer. AFAIK If that doesn't auto drop the current will stay higher too - if it auto drops it still needs to be explicitly reset to get over 120 MHz clocking.
I also backed Kris/onehorse on his KS to support him - but not yet seen a link to his KS posted - Paul's was the first note AFAIK. The link posted above was in answer to a question about Arduino on STM? I've seen it done before where non Teensy users get told where to go - that isn't SPAM.
I ordered a T_3.6 MPU9250 board from Tindie just to see if I could figure out how Kris gets the T_3.6 to run as slow as the low power STM with near the speed of a T_3.2.
Between WFE and yield() in this thread and the ZILCH multitask thread that uses the same or fewer cycles in task switching by replacing the default that eats those cycles in yield() scanning all Serial# ports for serialEvent() calls I've learned some things about Teensy. This thread seems to have had a reason to start - where it went from not being ignored has some value in Teensy usage and perhaps areas to revisit for improved Teensy use cases.
<edit>:Just searched LadyBug - found one each in this thread and the MPU9250 thread - only link is to OSH PCB.
<edit2>: BTW : I also abused the kinetis_hsrun_disable() with my WFE sample and current dropped lower - but then USB comms were spotty as my simple example did it - mostly sleeping at low speed: the only waking interrupt I had was the systick, and time got 2::1 dilated going 240 to 120 MHz. I wondered if duff's Snooze speed mode change dropped the hsrun (by default or with code) - but didn't see an answer. AFAIK If that doesn't auto drop the current will stay higher too - if it auto drops it still needs to be explicitly reset to get over 120 MHz clocking.
Last edited:
