Based on measures, I only see 1% more energy when using javascript. But the comparison is not conclusive because about 99% of the energy is used in the LED blinking itself and only about 1% used in the MCU. With my assumption I can conclude JerryScript cost twice power.
Interpreted code costing more power than native code? Maybe, but does it matter? Power profile of running blinky on STM32F4 in native code and in JavaScript... Can you even spot the difference? (note: @ArmMbed automatically puts MCU in Stop mode when idle)pic.twitter.com/Z0CzKGqN6C
-
-
-
Sure, and that's my point. While a 2x increase in compute power might seem a lot, it might only be 1% of your energy budget as the MCU is in deep sleep 95% of the time anyway and peripherals draw much more power.
Kraj razgovora
Novi razgovor -
-
-
Try to program the low power timer and let the CPU sleep in JavaScript
-
Well, that is happening automatically in Mbed OS already. All threads idle? Go to sleep. And the JS event loop is tied in with Mbed OS C++ EventQueue, which fits that behavior.
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.