I checked it out. It has an ARM version which is cheaper and alternative to RPi Wireless. RISC-V is still in its infancy, lack of GPU, only one core, more bugs... But it is a good starting point.
I have this version of Mango Pi. I run a port of Armbian (yes ARMbian). This board is good for trying out and testing, but not a deployment. The D1 is quite a ropey implementation of RISC V.
I think a major omission in this video are the nuances where the D1 really shines; I have used several D1-based boards, and there are a few situations where I would prefer a D1 over the Pi Zero. The first one is power consumption - performance is similar to an OG Raspberry Pi, but power consumption is scarily low. When running the Hercules-390 emulator with MVS 3.8j and several clients connected to WiFi and SD I/O happening, it was consuming less than 100mA @ 5V. This makes this an excellent candidate for situations where power is constrained or you need to get the most out of a battery. Another good thing about this chip. Another big strength of the D1 chip is that it has a variant (D1s) that comes in an (el)QFP128 package with 64MB of DRAM included in the package. This chip is perfectly doable to hand-solder since you don't have to deal with SMD, and it needs almost no support components besides SPI flash for your small Linux-based image. This effectively allows you to build your own Linux-based devices with average-ish soldering skills.
There are more OS available than you have mentioned and tested. You need to keep an eye on the gh project _Awesome Mango Pi MQ Pro_ by _boosterl_ which tracks a detailed list of working OS's and variants (there are 12 at time of post, both desktop and headless).
Given that this might be a good starting point for RISC-V development, how about a bare-metal tutorial on developing programmes (without an O/S) using the available toolchain(s)?
If you are looking for bare metal development in RISC-V , Prof. David Harris lectures in Computer Architecture on edx are a great resource. It uses the popular Sparkfun Red-V board which has a RISC-V core.
Nice to see you look at MangoPi! I have an MQ-Quad I got a few months ago and will be very curious to see what you think of it. It is unfortunate it doesn't have any MIPI connectivity, as it would be a good board to pair with a camera. I found it to perform pretty well, but video output at 4K was unusably broken. At lower resolutions it was okay, save for frequent mouse flicker. It runs warm, so the heatsink case is pretty essential.
there is better video drivers on some OS, but they also suffer from other issues. ATM it is still better to _mix'n'match_ your own OS from what is available - BTW the D1 Nehza is never going to be a "graphics behemoth" without a 3D assist, so why try .. if you want RISC-V _AND_ Graphics, pre-order one of the new Lichee RISC-V SoM or get one of the new SiFive SBC (if you can find them) - both are quad core, with the Lichee having upto 16Gb RAM
Thanks for your review ... I found that only the outside USB-C powered the board and not the inner one. I loaded Ubuntu Server and it is helpful if you have a USB to Ethernet adaptor, otherwise you need to download and manually install the wifi drivers and configure the network (text file). A single processor is restrictive, but as a new riscV board one this is a great start.
The file Compression probably uses operations, not realized on the RISC CPU Still a great alternative and a nice look at the architecture of the future :)
Yes I agree. In most deployments headless systems would be used but graphics processors can accelerate other tasks. Also as a comparison to RPi Zero I think it's important to mention!
@@LearnEmbeddedSystems well, being an embedded engineer... There are MCUs with dedicated peripheral ports for LCD panels, it's just not meant for xorg or Wayland or miir whatever their doing these days ..
For those that need driving existing displays, using devices different than RPi present a real challenge in term of drivers/libraries availability for non RPi. Although some of them are open-source, porting then to non-RPi increase dev costs due to having to allocate time and resources for the port. This is where the RPi huge community makes a difference.
I'm sure this has been clarified, but it's a Single Board Computer. You can see this like a motherboard with all the peripherals on single board for desktop applications unlike an embedded application.
It is an SBC, you don’t typically use them with arduino, they run their own operating system, hence you could see it had a full desktop loaded (badly) and the Ubuntu was accessed through a terminal. So you typically do development on the device itself although it is possible to cross compile on another system or you can probably remote access it using vscode.
This IS worth your money over an RPi by the sheer value of not being a closed-source abomination that boots the CPU using unknown code in the GPU first.
From the Ubuntu server it looks like they are just copying sipeeds RV Dock distro, I’m not sure how similar the boards are, they both use the D1 but might there not be differences to do with external components and drivers, etc?
AN Orange Pi Zero 2 is more costeffective and has lots of availability, RISC has a lot of ground to cover to get even as much software as the Orange Pi never mind the Raspberry Pi. Its ok if you want to experiment with RISC but its not for me!
These are my needs: to always be in stock, no geo restrictions, cost for delivery to be max 10$, to be GNU/Linux compatible, have GPIO + analogue pins and PWD, price to be extremely low to 35$ max. And I don't care about the brand :D or processor type.
@@elijah_9392 reduced instruction set, where as arm is more like cisc probably, complex instruction set, intel/amd are hybrids, risc/cisc combo. in lego terms risc has 10 blocks and cisc has a 100 blocks.
@@elijah_9392 I used to use RISC machines a lot in the 90's and 2000's. they were always far superior to any other machine on the market, ie IBM POWERPC beat every iteration of x86. the research and dev in RISC is much more mature than that of ARM
I’ve got a (literal) stack of Pi Zeros, so I have access to hardware if I need it. The real problem with Pi Zero is that it’s an obsolete platform. Running an ARMv6 core, it’s difficult to find usable software that still supports it. And being as complex as it is, it really needs a proper OS to boot. Unlike the Pi Pico which provides an embedded-focused ARMv6 core that runs more “bare metal”. Pi Pico is also readily available, which makes it a good option for new embedded projects. The Pi Zero 2 W is a good replacement on the higher end. At least on paper. In practice, I’ve never actually seen one. They began production right at the start of the supply-chain crisis. This made them so rare that only those willing to wait out pre-order timeframes have actually gotten their hands on one. Emphasis on *one* unit. Good luck getting them in any actual supply.
How is it so far ahead of arm? Risc-V is still a bit of a mess and I don’t think there are any Risc-V processors out there yet that compete with arm ones, Risc-V is generally slower so far and is no where near as well supported as ARM. Edit: Risc-V won’t necessarily be cheaper either, it is an open ISA not open hardware, the actual implementations of the processor can be closed source and in a lot of cases they are, so if you want to use Risc-V you either have to make your own or license a core from someone else. There may be open source cores available but the best ones will still need licensed.
so far there is no RISC-V product that truly out competes its ARM alternative - and it truly seriously lags in multi-core implementations. After all this time it's still a rude surprise to see how much the actual gap is between RISC-V vs ARM commercial implementations
@@TheSulross yeah people don't seem to see the flaws with risc-v they are just blinded by the fact it is an open ISA. They think it will be cheaper than arm due to no licensing costs for the ISA but you can keep the implementation closed source which is what many companies do. You will end up with the very basic RISC-V cores being open source but for advanced features or designs you will still need to license them from companies like arm. Just because the ISA is open doesn't mean the entirety of RISC-V is open source, it is literally just the ISA that needs to be open source. I have also not really seen any price difference between RISC-V and arm other than risc-V tends to be more expensive for less performance which goes against the common belief that RISC-V is cheaper than arm. With arm they set the ISA and there are rules on what modifications other companies are allowed to implement or how the implement it. With RISC-V there isn't any hard rules like that, I can create any non standard implementation or extension I want and it doesn't need to be compatible with any other implementations, which is a big problem with compatibility between brands, there is no way to stop them taking the ISA and doing whatever they want, at least with arm, arm has at least some say in the final implementation. RISC-V is a start and it is interesting and we will see a growth in it but I think it has far too many flaws to become any better than any other ISA.
@@gustavrsh You are very short-sighted. Now is not important. You failed to realized that America and Europe are not the future and that BRICS can only use Risc-V.
What exactly needs so much RAM? In general, RISC architectures need more RAM for code space than CISC ones. 64-bit architectures similarly need more code space than 32-bit architectures. Linux needs much more RAM than a bare metal OS. However, the real consumer of RAM is applications and their data. You choose the amount of RAM you need based mainly on that.