This started because I was interested in how the new ESP32 compared against other common Arduino compatible development boards.
Source code Arduino sketch on Github The source code is based on this post
Direct link to Google Docs Spreadsheet – same as embedded above – feel free to make a copy of the spreadsheet and link it below in the comments with new tested cores
Current measurement is for entire dev board so take it with a grain of salt – maybe someday I will go through and isolate just the SOC current draw for these devices which would be much more accurate.
[rpt name=”arduino-dev-boards”]
Interesting info, thanks. It would be awesome to include Microchip’s PIC32MX and PIC32MZ, also dsPIC33EP’s
If you have the ability to test those chips you can request edit privilege’s to this chart and add the info, I don’t have them easily available.
The Teensy 3.6 can be run up to 240 MHz with an “overclock” setting at the bottom of the menu in the Arduino IDE. I’d like to see your tests on that to see how it compares.
Hi Tim, I don’t actually have a Teensy 3.6 to test, that data came from the original link to the code source, if I remember correctly.. If you have one on hand you can download the sketch and test it yourself, I’m happy to add in more data to this spreadsheet if you do run the tests.
have you considered the possibility of overclocking the stm32f103 board (blue pill) now available on latest arduino?
Hi Jamie, I do not currently have that board, those original tests were from the previous user of this code.. If you have the ability to test and share results I will add them 🙂
Hi, it would be nice to include the new Adafruit ESP32 Feather board.
Hi,
It should be directly comparable to any of the ESP32 tests we have, the only thing that has really changed I believe is the code base that is being constantly updated by Espressif. But if the hardware is considerably different let us know 🙂 or post your own test, you can request edit access to the google docs.
Please don’t put units in the column data, but instead it should be in the column header. No charts or calculations can be done this way.
Results for ESP8266 divide long/multiply long is wrong. In reality divide long on ESP8266/80mhz is about 3-4us. Check the source and/or gcc optimization parameters.