threaded=True is slow

[silverchen0402]

Hi

I use your great module for my nodemcu-32s v1.1 board, micropython version 1.9.4, everything looks fine, but when I start webserver with threaded=True, the response is slow, about 4s, however it's ~200ms when threaded mode is False.

I have no idea why this happens, my micropython is clean, no other modules and I use IP to connect the device, so it seems not an DNS issue, could you give me some hints about this?

Thanks a lot.
-Silver

[jczic]

Hello and sorry for my very late answer!
It is strange indeed..
Just, can you try on your platform to increase the stack size before creating the threaded server with a call to _thread.stack_size([size]) ?

[jczic]

microWebSrv uses just one synchronous pipe-line (threaded or not) and more for websockets connections.
I'll remake it with my async lib : https://github.com/jczic/XAsyncSockets :)

[gengshenghong]

Thank you for your nice work.

Same issue here, it takes 3 seconds to respond with very simple http server (same with #42). I can confirm it works with "mws.Start(threaded=False)"

Suggestion:
Document it clearly about the limitation or by default set it as False at least. For first-time users, this waste some time as we usually use "default configuration" for testing purpose, and I started to think what is wrong with my network (as I cannot imagine a http test request takes 3 secs!). Of course, if it can be root-cased and fixed, that is better. :) Looking forward to the synchronous version.

Thanks,
Shenghong

[jczic]

Ok thanks.
I will test further because it doesn't make sense for a single thread on the server to take so long.

[jczic]

So, I have removed the thread allocation testing :
73cd549
But on computer or my emmbeded controller, the thread mode is faster than the no-threaded.
Strange ?

[gengshenghong]

Sorry, I was using Lyrat which has psRAM, but running normal micro-python. It seems there are some issues for normal micro-python on Lyrat (with psRAM) and there is a special version for it. Whatever, my actual intention is to use esp32 WROOM (without psRAM), on this board, it seems OK with or without threads, I'll focus on this board and test further.

Thanks,
Shenghong

[gengshenghong]

@jczic some updates, please take a look at micropython/micropython#4372

Now I focus on esp32 WROOM (no psRAM) module and take some careful tests.
I got same results using MicroWebSrv for the testing. It seems like thread behavior is strange in micropython....

[jczic]

Hello @gengshenghong, some news about this problem with readline IO ?

[jczic]

Hello, I released a fully new version (v2.0) of my web server here : github.com/jczic/MicroWebSrv2. Open source MIT, fully asynchronous, more robust, more fast and more efficient! It is delivered with a good documentation. Thank you for your support and feedback. ☺️ Jean-Christophe Bos Le ven. 21 déc. 2018 à 08:41, Shenghong <[email protected]> a écrit :

…

@jczic <https://github.com/jczic> some updates, please take a look at micropython/micropython#4372 <micropython/micropython#4372> Now I focus on esp32 WROOM (no psRAM) module and take some careful tests. I got same results using MicroWebSrv for the testing. It seems like thread behavior is strange in micropython.... — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#46 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAegLFdO56AedkbLVNjgHs5gDY6NlE_Bks5u7JCXgaJpZM4X-R9z> .

-- *Sincèrement,*