ESP8266 Community Forum

- Thu Aug 27, 2015 11:11 am #27179

What I mean is that you can't really expect NANOsecond operation in software. If only because of several things you cannot predict:

- cache behaviour
- interrupts
- compiler optimisation

Every interrupt can mess up the cache completely, for instance. If that means your code runs from flash instead of sram (cache), that will make a whole lot of difference.

NMI strobe ~38 ns (~6 CLK CPU 160 MHz)
Code IRAM section.
Synchronise fifo bus...
Output i/o pin = 10 ns jitter.
Min step ~100..200 ns. Steps only synchronise I/O bus

PWM and WLAN operation also use NMI...

- Fri Aug 28, 2015 5:26 pm #27264

PWM and WLAN operation also use NMI...

PWM = Hard GPIO_SIGMA_DELTA
WLAN no delays and not uses _NMIExceptionVector

- Fri Aug 28, 2015 5:30 pm #27265

PWM and WLAN operation also use NMI...

PWM = Hard GPIO_SIGMA_DELTA
WLAN no delays and not uses _NMIExceptionVector

Who is able to use - everything works, those who are looking for excuses - nothing works.
NMI, TIMER0 - 5us Call timer0_tst_is(). WiFi AP+ST On. WEB TCP GET/send file:

Code: Select all3:08:05.870> srv[80] 192.168.1.2:48733 [3] listen
3:08:05.870> srv[80] 192.168.1.2:48733 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:05.870> srv[80] 192.168.1.2:48731 [3] disconnect
3:08:06.008> srv[80] 192.168.1.2:48734 [3] listen
3:08:06.008> srv[80] 192.168.1.2:48734 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.008> srv[80] 192.168.1.2:48732 [3] disconnect
3:08:06.076> srv[80] 192.168.1.2:48735 [3] listen
3:08:06.134> srv[80] 192.168.1.2:48735 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.134> srv[80] 192.168.1.2:48733 [3] disconnect
3:08:06.258> srv[80] 192.168.1.2:48736 [3] listen
3:08:06.258> srv[80] 192.168.1.2:48736 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.258> srv[80] 192.168.1.2:48734 [3] disconnect
3:08:06.350> srv[80] 192.168.1.2:48737 [3] listen
3:08:06.350> srv[80] 192.168.1.2:48737 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.350> srv[80] 192.168.1.2:48735 [3] disconnect
3:08:06.520> srv[80] 192.168.1.2:48738 [3] listen
3:08:06.520> srv[80] 192.168.1.2:48738 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.520> srv[80] 192.168.1.2:48736 [3] disconnect
3:08:06.605> srv[80] 192.168.1.2:48739 [3] listen
3:08:06.605> srv[80] 192.168.1.2:48739 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.605> srv[80] 192.168.1.2:48737 [3] disconnect
3:08:06.697> srv[80] 192.168.1.2:48740 [3] listen
3:08:06.697> srv[80] 192.168.1.2:48740 [3] read: 348 of1[heap.xml] GET f[/heap.xml] head[200]:200 send: cf1 63 dis
3:08:06.697> srv[80] 192.168.1.2:48738 [3] disconnect

Code: Select allvoid timer0_tst_isr(void *arg)
{
   GPIO_OUT = GPIO_OUT^1;
   GPIO_OUT = GPIO_OUT^1;
}

Attachments

- Sat Aug 29, 2015 9:30 am #27321 No sigma-delta is something else. Is similar but not the same. I cannot give the same resolution as the software (!) PWM implementation, which, uses NMI, as claimed by Espressif themselves.

Wlan using NMI is also something Espressif claimed themselves. Don't shoot the messenger...