Workaround for 100% CPU usage problem on the virtual device

[sergeuz]

This PR fixes #1036 by suspending main thread for some minimum time on every Nth iteration of the application loop. That's probably not the best solution possible, but it allowed to reduce CPU usage from 100 to 10-15% on my machine (Intel Core i7), while keeping reasonable performance comparing to the real devices.

Note that calling std::this_thread::yield() has appeared to be insufficient to workaround the problem, also a simple unconditional sleep on each iteration makes an application significantly slower comparing to the same application running on the real device (at least 7 times slower in my tests).

Below is the simple test application that can be used to check how many times application loop is being invoked per second:

#include "application.h"

SYSTEM_MODE(MANUAL)

SerialLogHandler logHandler(LOG_LEVEL_WARN, {
    { "app", LOG_LEVEL_ALL }
});

uint32_t t = 0;
unsigned n = 0;

void setup() {
    t = millis();
}

void loop() {
    ++n;
    if (millis() - t >= 1000) {
        Log.info("%u", n);
        t = millis();
        n = 0;
    }
}

---

Doneness:

• Contributor has signed CLA
• Problem and Solution clearly stated
• Code peer reviewed
• API tests compiled
• Run unit/integration/application tests on device
• Add documentation
• Add to CHANGELOG.md after merging (add links to docs and issues)

Internal

• [GCC Virtual Device] workaround for 100% CPU usage problem. [PR #1167] [Fixes #1036]

[monkbroc]

Nice solution for this.

[m-mcgowan]

Great writeup and good work around for the time being. 👍

So...can we remove the hard-coded value of 25 loops and push that into a define? then we have a way of tweaking the number for different use cases.