I am a bit dismayed at this review being so unspecific, though. Which ESP32 chips is this competing against? Are these a bunch of first gen ESP32's (2016) we're running off against? Seems so. What about the newer -S2 or -S3 variants: might they have better power efficiency figures than the original ESP32? Hackaday has done reviews of the newer chips, but never compared their power. This ESP32-C3 preview is the first review that shows any power consumption figures (that I've found so far). Part of me hopes we find that this modest boost in efficiency is indeed a good win (notably for those with not a big amount of computational needs) but another part of me is hoping that this coverage has failed to show progress that has been made, that the better more modern chips it's not covering have already eeked out some better efficiency than what 2016 had brought us, and that this coverage has simply failed to look at more contemporary Espressif chips.
Anyhow, CoreMarks/mW are not the end all be all of chip efficiencies. It's a pretty narrow window to look through. This is clearly a smaller chip than the ESP32 & typically small chips get trounces in terms of power-efficiency, & while not winning it's clearly quite an improvement, holding up well. But there's still so many other factors that could affect real world power consumption. There's almost no other chips on the market that boast Bluetooth & wifi, pickings are so slim, Espressif is one of the very very few that have served this critical core innovation realm, the wireless maker & product creator. Any bump would help, and there's clearly a decent one here for the low end. But exactly what kind of gains we're really going to see, how much better power consumption we get, especially with more wake/work/sleep intensive cycles (versus the more work-oriented cycles demoed here), is going to be a critical determinant to what kinds of devices/products we all can make. What products we can fit in a power budget.
Basically the hope is that comments like this might be somewhat addressed:
> Interested to see how it does for power consumption- which always seems the achilles heal of the ESP32. Whilst the ESP32 has a good deep sleep mode, wakeup takes so long that it eats lots of power and saps how long a battery-powered project can run.
Also: contrary to the headline @victor82 gave: no, it's not hitting the market... yet. This is an early demo. Hard to guess how many months we have to wait until there is any kind of availability at all.
Well.. only if it's made on a more advanced technology. See, the RAM on an ESP32 actually takes near 85% of the total die, see https://zeptobars.com/en/read/Espressif-ESP32-Wi-Fi-Bluetoot.... The remaining 15% are for the radio, the peripherals and the cores. As there are 2 cores, each possibly takes less than 1% of the die.
