@jonbro oh absolutely. once those have been out for a few years and the firmware looks stable, i'm totally gonna put together like a v2 tangara mainboard with one
@jacqueline lol, when you made the announcement for the player, I looked at the design and went "huh, why isn't this a v3s, it's such a perfect fit"... especially as these chips can be made to be super power efficient, even while running.
... but I thought you must've seen those already 😬
that's like... 2-3 seconds from power on, to our firmware running (showing the splash screen is the first thing we do)
then a couple more seconds to have an actually usable device (we can improve that second gap i think; haven't looked much into it yet)
the number one issue with 'more powerful' SoCs people have yelled at me is that most of them are still fucking around in u-boot or whatever by this point.
@jonbro we show our splash ~100ms into our application firmware, so it must be the first and second stage esp32 bootloaders.
so initial hardware setup, loading the application into memory, that sort of thing. not sure what the headroom is for improvement there, but i assume it's small? would love to be wrong of course!
our esp32 boots fast enough that, when you're not actively using your device, we can just turn the whole thing off. so long as we restore your playlist and such when we boot back up, it's fine. it's not too annoying.
if your boot time is... what, 10 seconds? 20? then at some point you cross the "this is really annoying" threshold, and you can't do that anymore.
@colin_mcmillen mostly turning off a bunch of dev features like INFO logs and heap corruption detection. i might turn some of it back on (memtest on startup is surely a good idea?) but a lot of it is probably not really needed for release builds
Add comment