ESP32-S2 toolchain installed. Works, suggestions to make it clearer for the user of existing IDF

[rudi ;-)]

This is the version :
xtensa-esp32s2-elf-gcc (crosstool-NG esp32s2-dev-4-g3a626e9) 8.2.0

Thank you Martin, this matched to the same i have.

What I really don't understand is why the version is relevant for you

it is fast described:
the linked doc ( ESP32S2 ) has a download toolchain and is named:
"xtensa-esp32-elf-win32-esp32-2019r1-8.2.0.zip"

i was unsure, cause this link is 404 and is little different named to the latest toolchain for esp32
"xtensa-esp32-elf-gcc8_2_0-esp32-2019r1-win32.zip"

we got in the forum post the ESP32S2 toolchain link named:
"xtensa-esp32s2-elf-gcc8_2_0-esp32s2-dev-4-g3a626e-win32.zip"

in the hello_world esp32s2 make build but the "makefile" was saying, he want see
esp32-2019r1, but the toolchain is esp32s2-dev-4-g3a626e,
i asked then you for compare the toolchain version,
i knowed you use not the scratch steps and you was using the installer / linux
so i asked for the version you have in your succesfully installed compile build
that i can compare with the mine for this try,
i thinked you perhabs have an esp32-2019r1 likewise version
and my version esp32s2-dev-4-g3a626e is not the right perhabs, nothing more.

i got an answere from espressif that the esp32s2 toolchain version esp32s2-dev-4-g3a626e i use is right
and the named esp32-2019r1 is for ESP32, and this was in a GNU "makefile" checked,
but is not used in the cmake build cause cmake is used and not makefile
but i was warned while i try to build with GNU Make,
but do not disturb with cmake, cause using cmake and right toolchain

and in the second why in this ESP32 post where I was trying to make another point?

cause you wrote, you have installed ESP32-S2 toolchain and it works
i did all do in an other way ( windows, from scratch, gnu make ..) and was unsure to have the
right toolchain version, as the warning comes, i use different toolchain version.
so i did ask you here for the used version and pleased you to test the toolchain version
for compare with my used toolchain. nothing more we have the same version in use

so i know, if i generate an example, and my compile works,
you can test too - if you want - the example and can compile with same version
and only you can then upload to the esp32s2 ( MARLIN3 ) modul cause i have no peace of ESP32S2, CHIP7, MARLIN3 or other ESP32Sx
( yes - i have nothing this times - after 1000times pleased "add me to the list" i got nothing )
and i can not test the compilat on an real HW, only you
i must wait for the officially sale version - i am not on the beta list,
i feel like little boy crying under christmas like @martinayotte
you are lucky guy - i hope you know this Martin?

If you want to help, clone this, and let's pass it with more tests to ESP-IDF Framework instead of Arduino:
https://github.com/martinberlin/ESP32-benchmark

i will have a look into, sure, is it ESP32 or ESP32S2 Benchmark,
from name of repo, it is for ESP32 right? - do you will make Benchmark for ESP32?
perhabs a new post with Benchmark theme as headline?
perhabs in the show corner?
so your thread will not be alienated further more with more other testing points
and other question outside the posted theme from you.

best wishes
rudi

[fasani]

Nicht Böse gemeint!
Sure actually I want to do a full ESP32 test but that compiles on ESP-IDF, so nothing Arduino Framework, strictly ESP-IDF documentation and code examples. And it's no problem I will compile anything you want provided it compiles with IDF and share with you all results.

Arithmetic operations are not so interesting, since ESP32 is fast enough and probably the S2 will be the same in that area.
What I want to test is more in-deep operations, like:

- WiFi
- UDP traffic
- TCP, anything related to that communications part.
- Compression, https://github.com/martinberlin/Brotli I forked that so it works in ESP32

And at the same time measure consumption. I want to understand in what situations would be ideal to use S2 not only because has much more GPIOs but because it will consume less and be more efficient for the task.
A good example is described here: Ring Oscillator benchmark