Funding a FreeBSD Risc-V board, would you partecipate?

Would you crowfund a Risc-V SBC designed for FreeBSD?

  • Yes, let's rock!

    Votes: 7 50.0%

  • Nope, why?

    Votes: 7 50.0%
  • Total voters
    14
freezr

freezr

Hi folks,

FreeBSD panorama for ARM and RiscV board is very... meh... 😑
The majorities of these boards are built with Linux in mind and the support for FreeBSD is often very poor... 😭

However the FreeBSD Foundation may try to taste the ground and open a crowdfund to design and develop a RiscV SBC tailored for FreeBSD that works out-of-the-box from day one!

A crowdfund is a very good method to get something without risking too much, I am pretty sure it would be a great success!

I can already imagining a small sbc with two ethernet ports, 3 to 4 sata slot, 8GB minum of ram... 😁
 
Based on what CPU? Or do you plan to design the CPU too? One of the reasons why there are very few development boards available right now is the lack of CPUs (global chip-shortage).
And how much experience do you have designing boards like this? Or should someone else do the PCB design?

An SBC on a relatively small PCB means having 6 or 8 layers. That has it's own design problems (return signal ground planes). It's not as easy as it sounds.

freezr said:
I can already imagining a small sbc with two ethernet ports, 3 to 4 sata slot, 8GB minum of ram...
Click to expand...
One gigabit ethernet port, 2 SATA, maybe an M.2 SATA/NVMe slot, and 1 PCIe x4 slot. The PCIe x4 slot can be used for additional SAS/SATA controller or multiport ethernet cards making the board a little more versatile. Custom board size is tricky as that would require a custom case. So maybe aim for M-ITX, that way there's plenty of standard cases to chose from.
 
Why not cure cancer while we are at it? The Foundation's job is managing the project's finances, nothing else.

freezr said:
I can already imagining a small sbc with two ethernet ports, 3 to 4 sata slot, 8GB minum of ram... 😁
Click to expand...
And $600 per board due to low manufacturing volume.
 
SirDice said:
Based on what CPU? [...]
Click to expand...

I generated a misunderstanding... 😭

I'd like to have the FreeBSD Foundation sponsoring, promoting, making such project and I would be happy to crowdfund it!

I think is the best way to get a ready-made Riscv SBC!
 
shkhln said:
Why not cure cancer while we are at it? The Foundation's job is managing project's finances, nothing else.


And $600 per board due to low manufacturing volume.
Click to expand...

Looks compatible then... Board might pricey but mostly depends how successful the crowdfund is going to be...
 
SirDice said:
Yeah, that's going to be the biggest issue. Low volume also means expensive :(
Click to expand...
This is the real pain but you might mitigate it with feature...

For instance I have 3 boards, bought in different times, none of those can run FreeBSD but one (maybe, is the RPI2), eventually are $300 buck if you includes also: cases, PSU, usb dongle etc... Run all Linux (and worst Ubuntu) and are ARM based.

Clearly it should be reasonable affordable and having decent specs...
 
You wish is grateful to FreeBSD, and I also would buy such a board, if there were such (for a reasonable price, of course),

but as an experienced electronics design engineer who already did several mass production pcbs
I'd like to note
do not underestimate the number needed to be produced to get even in a range of a reasonable, comparable price, people will pay.

For something comparable like Raspberry Pi pico you'll need at least job lots of 30k...50k pieces for one production order.
If you want something 'fancy' like Pi 4 we're talking >100k...200k at least (many underestimate the prices of the connectors and plug sockets which are quickly the most expenisve parts.)

And talking cases:
The plastic doesn't cost much, but you have to get moulds for pressure casting made, which are app. around 20k...30k$/per piece.
You need at least two.
And also the mechanical design needs to be done, too, of course.

If there is a market for special FreeBSD single-board computers this could be done.
But unless you can ensure >200k buyers I'd recommend to check out the existing ones, and put effort in make FreeBSD run on them :cool:
I don't know - I hope I'm wrong :cool: - but I doubt there are so many FreeBSD users interested in using single-board computers.

You're not alone.
Just today HN published a link on that topic:
Installing FreeBSD on a Raspberry Pi

I don't know, if this helps - but I see less effort here instead of start a board.
 
Profighost the point is with the crowdfund you ask the market how much is willing before to get the product; then, If you meet the goal, you can start producing it besides that, to reply to covacat, you pay developers to write drivers for the board.
 
freezr said:
FreeBSD panorama for ARM and RiscV board is very... meh... 😑
The majorities of these boards are built with Linux in mind and the support for FreeBSD is often very poor... 😭
Click to expand...

Is this a software issue or a hardware issue?

I am not a developer (neither software nor hardware) but my preference would be FreeBSD stick to making great software to suit existing hardware than building hardware.
 
These are $250. They use COTS parts, don't need special drivers, run "standard" images and different OSes. No way anything esoteric like RiscV would cost $300 due to volume requirements and low demand. Getting anything to work with ARM is already kind of a pain in the rear and it's "popular."

You'd have to compile everything from source and it would have to be really, really fast with enough RAM and storage for compiling Rust/Clang/LLVM because I wouldn't want to set up a cross compiling toolchain just for one computer.

I wouldn't get one of these unless it was cheap and ran out of the box, even then I don't know what I would do with it, and there's no way it would be cheap.

Yes, that would be cool. It's a pipe dream, tho.
 
As I understand crowfunding it's a compromise between already done pre-development (you have to present something convincing) and gathering pre-paying customers.

If you got the money you do the final development - and the 'rest' needed to be done until series production is delivered to the customers.
All I'm saying is:
Don't underestimate that.
e.g. 50kUS$ are a sum most private people cannot afford to lose.
Within professional business, industrial production, 50k are nothing - not even peanuts.

Quickly things are overseen such as forgetting/underestimating product approvals like CE for european market, UL for USA,... and several others depending on your product, licenses...
you'd better think of them first before running later into:"damn, I didn't thought of that,...didn't knew this...SORRY FOLKS!!"

As far as I oberved things.
A) Crowdfunding is very good thing, I daresay brilliant.
B) Many crowfunding projects fail because of the things I mentioned were underestimated - set up (way) to optimistic
ending in crowd and project-initiator seperate not in good, often lost (much) money, suing.... - not good.

...it's also a good possibility to scam: "Here, great product. Invest!" - "sorry guys." (muharhar 😈)
But I assume that was not your intention.

So, bottom line - I say it again:
Having an own special FreeBSD board would be of course not a bad thing at all.
But I say:
Check the market carefully first to ensure you'll really get enough customers
and do a proper project plan
before any serious effort or even money is spent on it.

and as others already mentioned, maybe the (less) effort spent otherwise could lead to the same (quicker, or even any) solution.

peace out.
 
Just for giggles I checked Mouser, they don't have a single RISC-V CPU or MCU in their assortment. As far as I know only SiFive has a RISC-V CPU that could be used in a generic computer. All other RISC-V CPUs are mostly very limited microcontroller type CPUs (would lack the required features needed to be able to run a full OS on it).
 
at >= 4GB and decent I/O performance x86 based boards are cheaper and better supported
while having about the same power requirements
 
I see very few enthusiasm... Even though this is an hypothetical thought and everything can be made and requested, still for free... 😛
 
covacat said:
at >= 4GB and decent I/O performance x86 based boards are cheaper and better supported
while having about the same power requirements
Click to expand...
Sure. But I do like the open nature of RISC-V. And if I could find and buy an actual proper CPU I'm willing to invest some time in trying to figure out what would be needed in order to build a computer around it. Then also open source that design.
 
SirDice said:
Sure. But I do like the open nature of RISC-V. And if I could find and buy an actual proper CPU I'm willing to invest some time in trying to figure out what would be needed in order to build a computer around it. Then also open source that design.
Click to expand...

Eventually Risc-V will take over because open-source is more prolific than closed source development...
 
freezr said:
Risc-V will take over because open-source is more prolific than closed source development...
Click to expand...
Note that only the specifications are open and free to use. The actual implemented CPU isn't.
 
freezr said:
I see very few enthusiasm...
Click to expand...
In general, from a very high-level point of view, the BSD communities tend to be less "enthusiastic" compared to what one might know from similar Linux communities. This however does not mean that we wouldn't want this or wouldn't support it. But personally I prefer people to be realistic and then get a proper product out of this rather than being hyped, showing overly enthusiastic signs, somebody then commits to the idea and in the end it turns out that nobody really cared as much as initially demonstrated :)

I have to go with SirDice on this one: RISC-V is something I'd be willing to privately invest time & money in. But I'm not going to jump in the air and clap my hands just to make someone feel like there is interest :p
The hurdles are also non-trivial - which is again not to say that we (as a community) are opposed to any efforts into this direction.

In any case, any of your efforts will be greatly appreciated. No matter how far you get :)
 
Crowdfunding hardware is the worst.
They never deliver on time and there is no one to blame problems on.
Documentation and Support is usually not offered.

What happens in 2 years when RISCv666 hits? Do it all again?
 
Phishfry said:
Crowdfunding hardware is the worst.
They never deliver on time and there is no one to blame problems on.
Documentation and Support is usually not offered.

What happens in 2 years when RISCv666 hits? Do it all again?
Click to expand...

Making a hardware and manage a crowdfund campaign requires knowledge and expertise, I don't expect the first attempt is going to be perfect and accurate 100%, but I guess people behind FreeBSD are able to bring up a decent product, that maybe could encourage a second campaign. Anyway, anything of what I said would ever happen, so we continue to follow as usual: ugly hardware and awful *BSD support... 😞
 
I think the Arm maker board market is very fractured.
Some want purely embedded boards like BeagleBoneGreen some want a desktop with HDMI.

They try and jam all these features into a 100x75mm footprint for the least amount of money.
 
You must log in or register to reply here.
Back
Top