SoylentNews is people
SoylentNews
The Raspberry Pi Foundation's first microcontroller, the Raspberry Pi Pico is now on sale at $4. Raspberry Pi is normally associated with single board microcomputers. This microcontroller uses the RP2040 dual-core ARM Cortex-M0+ chip. The board has support for C, C++, and microPython.
We had three principal design goals for RP2040: high performance, particularly for integer workloads; flexible I/O, to allow us to talk to almost any external device; and of course, low cost, to eliminate barriers to entry. We ended up with an incredibly powerful little chip, cramming all this into a 7 × 7 mm QFN-56 package containing just two square millimetres of 40 nm silicon. RP2040 has:
- Dual-core Arm Cortex-M0+ @ 133MHz
- 264KB (remember kilobytes?[*]) of on-chip RAM
- Support for up to 16MB of off-chip Flash memory via dedicated QSPI bus
- DMA controller
- Interpolator and integer divider peripherals
- 30 GPIO pins, 4 of which can be used as analogue inputs
- 2 × UARTs, 2 × SPI controllers, and 2 × I2C controllers
- 16 × PWM channels
- 1 × USB 1.1 controller and PHY, with host and device support
- 8 × Raspberry Pi Programmable I/O (PIO) state machines
- USB mass-storage boot mode with UF2 support, for drag-and-drop programming
And this isn't just a powerful chip: it's designed to help you bring every last drop of that power to bear. With six independent banks of RAM, and a fully connected switch at the heart of its bus fabric, you can easily arrange for the cores and DMA engines to run in parallel without contention.
[*] By comparison, the Apple II computer (introduced in June 1977) had: 4-48 KiB of RAM, a 6502 processor (running at 1 MHz), and an Introductory price of US$1,298 (equivalent to $5,476 in 2019).
Additional coverage:
- Raspberry Pi Enters Microcontroller Game With $4 Pico
- Raspberry Pi Pico, RP2040 and Adafruit family of boards
- The Raspberry Pi Pico is a new $4 microcontroller
- Raspberry Pi Pico, the first microcontroller released by Raspberry Pi Foundation, based on the new RP2040 MCU.
- Raspberry Pi Pico microcontroller: specifications, features and RP2040
- Programming Raspberry Pi Pico with Python and MicroPython
- How to solder GPIO pin headers to Raspberry Pi Pico
- Raspberry Pi Pico FAQ
- Welcome Raspberry Pi to the world of microcontrollers
- Hello Raspberry Pi Pico and RP2040!
- Happy Pico/RP2040 Day!
(Score: 0) by Anonymous Coward on Friday January 22 2021, @02:15AM (9 children)
This seems a little imbalanced. 133Mhz, is 486DX4 territory, which you could do quite a lot with. But if I remember right, back in those days 4 - 16MB of RAM was a typical install. With a 4 to 1 disk cache, your still looking at just over a meg including user space applications. Even for a simple router (without firewall) that isn't much RAM. You couldn't even load a 360K FDD completely into memory.
Incidentally I had an apple II+ with 48k, and a commodore 128. But those were tiny, tiny kernels.
What could you actually use one of these for?
(Score: 4, Insightful) by takyon on Friday January 22 2021, @02:38AM (4 children)
It's a microcontroller. You want to step up to the Raspberry Pi Zero if you want an operating system.
Incidentally, this launch seems like good evidence that a revamp of Raspberry Pi Zero is coming at some point. They should switch to USB Type-C, DisplayPort over Type-C, 4 cores of some kind, wireless by default, and ditch the $5 price point.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Friday January 22 2021, @07:55AM (3 children)
Why? You have your RP 4 for 4 cores. Wireless is not by default because you don't need it for many workloads. I don't understand your comment at all.
The $5 price point is for non-networked zero.
(Score: 2) by takyon on Friday January 22 2021, @08:14AM (2 children)
http://linuxgizmos.com/rpi-zero-w-clone-offers-quad-core-power-for-15/ [linuxgizmos.com]
https://www.notebookcheck.net/Orange-Pi-Zero2-The-Raspberry-Pi-Zero-alternative-is-finally-orderable.501626.0.html [notebookcheck.net]
Why not? People run it as a desktop, media player, etc. Alternatives are going to quad-core (not the big cores of Pi4, obviously) at around $15.
I didn't specify a time frame for doing it. Raspberry Pi Zero will remain in production until at least January 2026. They could launch a new version in 2025.
Pico can take on SOME of the duties that Zero was used for. They can drop the non-wireless option and $5 price point for a Zero2.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Friday January 22 2021, @10:19AM (1 child)
RPi will always use HDMI as long as HDMI is the standard on TVs. It's important to them that a dedicated monitor is not required.
USB-C sure, it's a better physical connector than other USB ports and can supply more power, but DisplayPort over USB-C is kind of a loopy idea to begin with and only found on high end hardware. And DisplayPort just isn't required.
What this does is bring Raspberry Pi branding to a more powerful microcontroller than Arduino. The RPi has iffy GPIO and lots of CPU power, Arduino has good I/O but so little computation power that you can barely connect to networks with it. This will split the difference nicely.
And it opens up the possibility of a RPi with both on board, a real RPi system with a microcontroller on the side for realtime I/O. The Beaglebone can do this but it's expensive and outdated. That would be pretty darn handy.
(Score: 2) by takyon on Friday January 22 2021, @01:22PM
Fine, HDMI Alt Mode [hdmi.org] over USB Type-C.
RPi Zero has a mini-HDMI port. If they change it, it would be to micro-HDMI or USB Type-C. Choosing USB allows the port to be multi-use, choosing micro-HDMI allows reuse of cables that people bought for RPi 4B/5/???.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 4, Informative) by KilroySmith on Friday January 22 2021, @04:09AM
FreeRTOS would be an appropriate kernel for this type of device, with a footprint starting around 10KB depending on the services that you include. That won't include the kind of capabilities that you'd get in a Linux kernel, but if you need those kinds of capabilities I've heard that the Raspberry Pi organization might have a more appropriate board...
(Score: 2) by JoeMerchant on Friday January 22 2021, @02:04PM (2 children)
I have always been underwhelmed with onboard RAM provisions in microcontrollers. I took those theory classes where they taught the "Harvard Architecture" maxims of ROM to RAM ratios, but somehow my applications never fit within those norms and we almost always ended up installing a 32Kx8 RAM on our 8 bit micros.
These days, the question is: what libraries are you wanting to install? Is 264K enough to run lib-gcrypt?
Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
(Score: 2) by VLM on Saturday January 23 2021, @04:55PM (1 child)
Not disagreeing with anything, but single threaded apps have it easy, the real killer is when you got multiple thread RTOS and each thread gets spawned off with a dedicated statically defined stack.
So on a single thread app with one big shared stack things can get kinda sloshy and its OK if you can't do TCPIP stuff AND framebuffer graphics AND processing because being single threaded you can't do them all at the same time anyway... Multithreaded HURTS without enough memory.
My favorite thing is how some RTOS like ARM Mbed OS just go insane if you overflow the stack. Hilarious! Or sometimes they crash. It can be a mess.
I like using multithreaded RTOS to do stuff like decouple the scheduling of my IO, GUI, and processing, compared to doing in arduino-style in one big loop, but there is a memory cost...
(Score: 3, Interesting) by JoeMerchant on Saturday January 23 2021, @05:04PM
I have found that multithreaded RTOS work pretty well, as long as you are seriously underutilizing the capacity of the system, like 50% or less. When things get tight, prioritized interrupts and all the rest frequently throw hard to debug unpredictable behaviors into the mix and people really have a hard time setting them right.
When possible and it's not always possible with things like TCPIP and framebuffer and others happening simultaneously, but when possible I like to define a single monotonic timed interrupt for the system, like once per 500uS or whatever, and then run a deterministic choreographed script of what happens on each phase of that interrupt, so maybe do 40 cycles with 20 of them servicing things that need to happen every 1mS, 10 more things that need to happen every 2mS, then the remaining 10 can service things that need to happen within 20ms... put an "in interrupt service" signal on a GPIO line and watch it on a scope, or if you can dump the assembly and ensure that your code WILL finish within the allotted 500uS. Now, anything else that needs doing but isn't time critical can get serviced in the "main loop" that runs whenever the time critical stuff isn't happening. I've made several systems that way and they never had the "jeez, it only happens once every 6 weeks or so, but it's really weird I can't explain it" bugs.
Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end