here i’ll document my explorations in parallel computing, and other home electronics projects.
my dad’s a software engineer. he bought me this arduino / diy electronics set 10-15 years ago, hoping and praying that i would take interest in his field. unlucky for him, i was too adhd to focus time on those things, instead i just played runescape or whatever else 10-15yo me was doing. fast forward 10-15 years, i notice the bin in the corner of my eye and gears are turning. better late than never, right?
i have a system now that enables me to finally get to the things i want to do and be productive etc. i took out the arduino and started tinkering. never really done this shit before aside from a robotics class in highschool i took forever ago, and even then i didnt really have the attention span to get much out of it. (unfortunate, cuz the teacher was super super good). took me 2 whole hours just to be able to make an led blink. the issue was that the arduino lab micropython ide wasnt giving me enough feedback to diagnose the issue, and had to switch to something bulkier, the original arduino ide. and even then, i had to enable “verbose logging” or something like that so that it prints out pretty much everything. then we were good.
anyway turns out arduinos are super, super weak. the (link) arduino micro was trying to use, has as its main processor this “atmega32u4” chip, which if you check the specs of, can pretty much do nothing. and it costs more than a pi pico. which is a gazillion times better. (which turns out is also weaker than the cheaper esp32 which also comes with wifi, go figure.) i went down a big rabbit hole investigating all the limitations and stuff having virtually no experience in the department of home electronics. turns out the benefit of arduino is that its accessible to use as the front facing variables and functions you call look pretty by default, thanks to the work they do over at arduino the company.
in parallel, my gf was interested in parallel computing and/or doing a ml project, using opencv image processing tools. i had the genius idea to use this electro-whatever as motivation to loop her into this. what if we can buy some cheap parts like a camera module, esp32 whatevers, and some pi picos to try both of those? i was warned a bunch by other people like “its not gonna work, pi picos are too small you wont be able to fit kubernetes on it” and some other lame excuses. idc. when theres a will theres a way. if that package wont suit my needs, ill have to find a smaller one that does, or have to make it myself, which would be interesting in its own right. “nooo, even if you reinvent the wheel, there wont be enough processing power left to do anything.” sorry friends, i will be naive and cross that bridge when i get there. that may very well be the case, i certainly dont have the depth of understanding to gauge that possibility at this current state, but i think it would be a very interesting task to dig and really look at coding on the metal, to investigate and determine for sure the feasibility of this endeavor.
you know what? maybe the post format is good. maybe ill add another post when i get to progressing more on this.