#avr Logs

Dec 17 2018

02:00 AM nabil__ is now known as nabil
04:52 AM nohit: new series from ST https://www.st.com/en/microcontrollers/stm32g0-series.html?querycriteria=productId=SS1960
04:55 AM _ami_: "Setting the lowest ever price point, the STM32G0x0 Value Line makes no compromise on what matters" - that's interesting!
04:58 AM nohit: that's made to compete with the 8/16bit market
05:09 AM polprog: its 32 bit, right?
05:09 AM polprog: maybe ill order some..
05:13 AM nohit: yeah
08:15 AM vmt: I wonder where the price will settle. st's m0s are already quite cheap, have more functionality (minus eeprom) than attiny/mega, and are much cheaper than xmega alternativex
08:15 AM vmt: alternatives* even
08:20 AM nohit: https://www.st.com/en/microcontrollers/stm32g070rb.html
08:20 AM nohit: 78cents
08:20 AM nohit: and outperforms pretty much every AVR
08:21 AM nohit: 36K of ram for 78cents
08:24 AM nohit: this costs 1e and has 0.5K ram https://www.mouser.fi/datasheet/2/268/Atmel-2586-AVR-8-bit-Microcontroller-ATtiny25-ATti-1315542.pdf
08:32 AM nohit: this is close in spec but cost over 4 times more https://www.mouser.fi/ProductDetail/Microchip-Technology-Atmel/ATMEGA1284-AU?qs=sGAEpiMZZMuHCAZ7U3Ea2g%2fOz4B1Ivlx
08:52 AM rue_mohr: canada: plus $10 shipping
08:53 AM rue_mohr: the stm32 does cost less, and is more powerfull, but is more complex to code for
08:53 AM rue_mohr: i still find myself using avr for quick things
09:32 AM vmt: st's stdperiph is super easy to work with
10:12 AM nohit: stfperiph is legacy tho
10:12 AM nohit: *std
10:13 AM nohit: now you have HAL and LL drivers
10:13 AM nohit: or the cube framework/ecosystem
10:21 AM vmt: "legacy"
10:22 AM vmt: also cube generates HAL code, but there's no absolute need to use HAL unless you're on an m7(/+)
10:23 AM Ameisen: I haven't tested it yet
10:23 AM Ameisen: but apparently:
10:23 AM Ameisen: 1. Clang's __builtin_constant_p support now matches GCC
10:23 AM Ameisen: 2. Their AVR support has improved
10:23 AM Ameisen: I'm guessing they took some of my work
10:23 AM nohit: cube and hal are part of the same thing
10:23 AM vmt: now what did i just say?
10:24 AM nohit: you said "legacy" because you didnt want to admit that you are wrong
10:24 AM vmt: i said "legacy" because it's still completely usable and often times simpler to use than HAL
10:25 AM vmt: the exceptions being, there are no stdperiph libs for f7, or the g0 i figure.
10:25 AM vmt: though i guess the g0 availability is quite poor still, but go figure
10:26 AM Ameisen: Once you are at the point where an xmega is appealing... why not jump to arm?
10:26 AM vmt: there really is no reason not to
10:26 AM nohit: they just released AVR-killer
10:26 AM nohit: ST
10:26 AM Ameisen: heck, 90% of the uses for AVR overall are better suited with ARM
10:26 AM vmt: nohit: so want to continue arguing for the sake of arguing?
10:27 AM vmt: or are you asshurt by me disturbing the status quo over at #avrs?
10:27 AM Ameisen: lots of companies seem to sick wtih AVR due to momentumn
10:27 AM nohit: what :D
10:27 AM nohit: how am i arguing ?
10:27 AM vmt: i remember you, now have a good time being ignored
10:28 AM nohit: i was talking to Ameisen
10:28 AM vmt: idiots such as yourself are pretty much the main reason i left #avrs
10:28 AM nohit: and not a single fuck was given when you left
06:01 PM davor_ is now known as davor
06:58 PM rue_mohr: cehteh, ?
06:58 PM rue_mohr: did you get what you wanted?
06:58 PM cehteh: i dint care yet, was more like a quiz
06:58 PM rue_mohr: nohit, whats up?
06:59 PM rue_mohr: vmt, is avrs full of jerks?
06:59 PM cehteh: not completely trivial and either way leds have some tolerances, i rather pick resistors by experiment
07:00 PM cehteh: currently the 'yellow' is a bit too greenish etc
07:00 PM rue_mohr: oooh, I just made a tripple equation solver
07:00 PM rue_mohr: 3x3 simotanious equation solver
07:00 PM cehteh: on avr?
07:00 PM rue_mohr: hah no
07:00 PM rue_mohr: I could, the calculation might take a ms tho
07:01 PM cehteh: i had this fun here, used a sovler to generate the function to caclulate the results i need
07:01 PM cehteh: which was helluva complicated and had lots of 32 bit overflows
07:01 PM cehteh: now i settled just with bisecting the result :D
07:04 PM cehteh: seen my rant about my schoolbook divide implementation is *much* faster than what gcc did? :)
07:04 PM rue_mohr: no, but ok
07:04 PM cehteh: in C not asm not really optimized
07:04 PM rue_mohr: I think gcc uses the most universal method it can
07:05 PM cehteh: but inlined and 32/32=16bit ...
07:05 PM rue_mohr: shift subtract, shift subtract
07:06 PM cehteh: its runtime costs depending on the values, and where it matters most it gets faster .. perfect
07:06 PM rue_mohr: hmm, I need to work out which goals are important
07:06 PM rue_mohr: I have lots of goals
07:06 PM rue_mohr: "kill all humans"... and how did this one get in here anyhow?
07:07 PM rue_mohr: micro robot
07:08 PM rue_mohr: hmm
07:08 PM cehteh: haha i ordered stepper motors from amazon, with uln2003 driver board for 1.20Eur a piece
07:09 PM cehteh: but only when you order a single one, increase the number and it adds 9Eur shipping
07:09 PM cehteh: so i orderd one per day for the past days :D
11:31 PM day__ is now known as day