Balthazar Personal Computing Device, a 13” RISC-V laptop | Hacker News Reader

Balthazar Personal Computing Device, a 13” RISC-V laptop

balthazar.space

106 pointsSamWhited60 comments

In case it wasn't clear to anyone: this is a product very early in the planning phase, to the extent that many of the features are unspecified (no processor has been chosen, they aren't even certain on whether it'll be RISC-V or ARM), underspecified to the point of being unintelligible ("detachable USB gender-changer dongle"), or outright ridiculous (SSH-based communications between the keyboard and CPU).

I wouldn't hold my breath waiting for a product here. This does not have the smell of success on it.

Like seriously what would an ssh client running on the keyboard even achieve? Maybe they’re thinking of encrypting it. But keyloggers running at the OS level don’t care. Or is this to force people to only use their keyboards? Or maybe it’ll just add a billion to their SPAC buyout.

I guess you could have a wireless keyboard without some of the bluetooth security issues?

SSH would not be the way to accomplish that.

For anyone else who wants to take a peek for themselves at the concerns raised in this comment: https://webcache.googleusercontent.com/search?q=cache:https:...

Another thing that doesn't seem to line up with reality is the idea that they might convince Nvidia to make their stuff FOSS enough to be included in this project.

I assumed that when I read the words riscv and laptop, but then I saw nlnet was backing the project. I wonder what they are hoping to get out of the project.

> While being versatile and robust it also follows CERN OSHL, GNU-GPL, FOSS, EOMA, ISA and even CC guidelines

This is almost the most concerning part. They're treating FOSS like it's some kind of standard. It's not. FOSS just means software that is free and open source. Something is, or isn't FOSS. You don't follow some standard to deliver a FOSS product. I also couldn't find a single reference to CERN OSHL anywhere online other than on their website. The closest thing I could find was Out-of-school-hours learning... again, if that's what they meant this is doomed.

I think by "CERN OSHL" they mean the CERN Open Hardware License: https://en.wikipedia.org/wiki/CERN_Open_Hardware_License

There's no "Source" in the name.

the acronym salad seems to support it being vapourware, but I think CERN OSHL is a typo for a real thing: https://en.wikipedia.org/wiki/CERN_Open_Hardware_Licence

There's gotta be some reasoning behind those but when you put it that way it sounds like some hilarious marketing speak.

"We're putting blockchain technology in the trackpad"

"Machine learning in the power button"

This may sound insane, but machine learning in the power button may be a thing already. Fingerprinter readers (often built into the power button) have moved from “match on host” to “match on chip”, and it’s possible some of them use ML-based matching.

indeed, I believe the M1 mac power button is also the fingerprint sensor...

My 2019 MacBook Pro uses the power button for the fingerprint sensor. Apple has been doing that for a while, since the iPhone 6s as I recall.

"We used a malleable logarithmic casing to construct the chassis of the laptop"

Looks like the good ol' "Have you tried logarithms? " worked for them.

https://xkcd.com/451/

Note, the grandparent is not actually quoting the website, they are referencing the fictional Turboencabulator[1]

[1] https://en.wikipedia.org/wiki/Turboencabulator

"Includes a laptop bag made with Fine Corinthian Leather"

(due to supply constraints, laptop bag will be made with Nylon.)

Yeah - clicked on this expecting to see vaporware; was not disappointed.

Yes. While it sounds impressive, it seems more like a fantasy jumble of every desirable feature they can think of than an actual plan.

"Keyboard is QUERTY". How do you misspell "QWERTY"? Yet I'm hoping they did.

They seem to be hoping for a Pixel Qi style screen (but self-developed?), but describe its reflective mode as: "A non-reflective - a button driven backlight “off” monochrome mode for very low-power use in sunlight"

Confusion between trackpoint and trackball: "track-point ball"

Perhaps it’s just an English as a second language thing? Note, I have no idea the background of who is behind the project..

My favorite is the GPU paragraph:

> While NVIDIA would be technically possibly the best performance choice, issues arise with general attitude towards FOSS compatibilty. In any case we look and listen into NVIDIA and CUDA options, accelerating totally open documentation from NVIDIA's side. Also the NVIDIA's Jetson IoT and AI platform is researched for a compatibility with RISC-V architecture, ISA and FOSS

Uh, surely they don't mean the retro ISA bus, right? :D

On a more serious note if I could build a big (dGPU oriented) open laptop I would put an MXM slot in there, so that existing Radeon cards could be used.

ISA == Intruction Set Architecture.

Of course, but the way they use it makes no sense.

There's another level of WTF in that paragraph. "NVIDIA Jetson" is a marketing name used for a series of Tegra development boards. Tegra is an ARM SoC in and of itself. It isn't a discrete graphics solution.

What, exactly, is unintelligible about a "detachable USB gender-changer dongle"? A gender-changing adapter is one that has either a male connector at both ends or a female connector at both ends. They're used in damned near anything involving plugs and cables - or hoses, for that matter. It's not a mystery marketing term or anything. In this case, it would be female at both ends, providing a degree of physical security when not in place (as both the USB device to be plugged in and the dongle-less computer would have male connectors), and, as described, would also act as a security dongle when attached (the USB port needs to see the dongle and have it password-activated in order to use the USB port, so a simple passthru gender-changer won't give access).

Any fan of '90s JRPG's will tell you this product shares its name with an engineer who finished his masterpiece in the year 2300.

The irony is strong.

Looks like vapor.

The MNT Reform 2 [0] already exists, shares common goals, and is entirely compatible with a future risc-v SoM with its modular design.

[0] https://mntmn.com/reform

I'd happily buy a RISC-V laptop, but I think it's too early, mobile phones and servers will come first. Also the artist rendering is ugly, I would expect an ultra-light laptop from a low power architecture.

I want to see how it competes with Apple's M1 processor in speed (especially that the web page writes that it's a fast laptop) and power utilization.

I think risc-v might work best as a server

For phone/desktop/laptop a rough spot would be the GPU. And possibly perf/watt.

The M1 meanwhile most likely has billions of dollars of development behind it.

I don't see why RISC-V couldn't have billions of dollars behind it in 10-20 years.

For Chinese phones I don't see any better architecture right now, as they can't depend on ARM forever, it's too risky for them.

Isn't China pretty heavily invested into MIPS with the Loongson? Is there any good reason for them to drop that for RISC-V?

RISC-V has a huge effort behind it to run all available open source software. Even though there’s a MIPS Android port, it just doesn’t have the amount of backing that RISC-V is getting from hundreds of companies (basically all big companies)

I'd expect less activity around the MIPS port because it is, by and large, already fully functional and a first class citizen of Android, while RISC-V is still being fully bootstrapped.

Further, software like Chrome/Chromium and V8 already have functioning MIPS ports, which will be (a lot of) additional work not currently being performed for RISC-V.

Compared to RISC-V, almost everything in the Linux ecosystem should already "just work" on MIPS. I think it's fair to ask why there wouldn't be more development of MIPS, which is effectively a Chinese owned ISA at this point, since it's more mature than RISC-V, if not as "exciting".

To quote Elon Musk: ,,The most entertaining outcome is the most likely''.

I see just the opposite when I look around Reddit: MIPS is dying. Google Play services is not updated on it, and I don't see why people would port new games on it.

You need neural networks running efficiently to run all the face swap algorithms (V in RISC-V was for vector support originally), GPUs supported (again a RISC-V GPU ISA extension is in development), JIT acceleration (J extension). If it's not exciting for developers, it won't win for consumer devices.

I bought a ridiculously cheap Android tablet from AliExpress many years ago, and didn’t even know that Android on MIPS was a thing. Until I received the tablet in the mail and learned that it had a MIPS ISA CPU. Many apps on Google Play would not install on the device, because they were for ARM only. Things may have changed since then idk, I haven’t used Android in many years. But the experience at the time was, well... The tablet ended up in a drawer and has been sitting there since.

I should add though, regardless of whether app compatibility is a problem or not still that app compatibility was not the main reason for relegating the tablet to sit in a drawer.

The main reason was battery life. Not surprising given how cheap it was. And it was really slow. But those two things aren’t meant as a knock on MIPS. They obviously put in the cheapest CPU and the lowest capacity cheapest battery to make the tablet as cheap as it was. Just adding this to explain why I ended up not using the tablet for anything at all.

Basically Android is all about ARM nowadays, the only place where alternative CPUs matter is running on Chromebooks using Intel CPUs, and even there Google has a ARM JIT for dynamic translation for APKs that are ARM only.

MIPS and Intel CPUs on Android phones and tablets are dead.

MIPS is no longer relevant on Android, it is even no longer supported on recent NDKs.

The last I heard of Loongson they moved to a custom MIPSish ISA as the native one, with in-processor support for MIPS, RISC-V, ARM and x86.

https://www.zhihu.com/question/414069789

I thought China abandoned MIPS for PowerPC, then abandoned PowerPC for ARM, then abandoned ARM for RISC-V.

> I want to see how it competes with Apple's M1 processor in speed (especially that the web page writes that it's a fast laptop)

That's a goal, not a specification. The project hasn't chosen a processor. They aren't even sure it'll be RISC-V.

Maybe the RISC-V people will now move on to cloning the M1?

I don't think anyone has even designed anything in the same ballpark as M1 in transistor count let alone built one for RISC-V.

Sure, in speed nothing compares to M1, but I have heared amazing claims for energy efficiency even with much worse process nodes.

From the description it sounds like this will effectively be an SBC in a laptop case. Is anyone aware of such a thing? I would certainly be keen on a laptop "enclosure" for an SBC.

EDIT: found a bunch here, but they feel a bit clunky to my eyes: https://magpi.raspberrypi.org/articles/the-best-raspberry-pi...

It's getting hugged to death.

Mirror: https://web.archive.org/web/20210205204013/https://balthazar...

Can it possibly be be used with Libreboot? It would be nice to have some newly manufactured hardware in laptop form that Libreboot can use.

Can libreboot target any ARM SoCs? It seems like u-boot with a verified boot path is not a common configuration, while the coreboot ecosystem has a number of security oriented payloads.

The Libreboot wiki lists one ARM based system as supported, a particular Chromeboook:

https://libreboot.org/docs/hardware/c201.html

I think it's fair to point out libreboot is a fork off coreboot, and while libreboot is basically a dead project at this point, coreboot development continues. Chromebooks tend to use coreboot natively, and on some chromebooks, you can recompile and reflash coreboot, sometimes with no binary blobs.

Google makes this (reflashing the firmware) surprisingly easy in many cases, although there are often many caveats to running linux/alternative firmware on chromebooks, to the point it's kind of hard to recommend buying one for this purpose.

> Can it possibly be be used with Libreboot?

No. Libreboot lacks support for any of the hardware under consideration by this project.

Libreboot is essentially a dead project. It only supports a limited number of Penryn-era (roughly 2007-2010) Intel laptops, and -- strangely -- the RK3288C-based Chromebook C210.

That’s a bit harsh. You can still buy coreboot machines:

https://www.coreboot.org/users.html

The projects are related, at least.

That's because coreboot is an active project, and is still being updated to add support for newer platforms. libreboot isn't.

The concept image indicates a red trackpoint at the center of the keyboard. Hopefully that makes it into the final product.

They sell a $280 Windows 10 Pro laptop in India that is not slow at all and has an ethernet port/hdmi port etc. I've found it much more useful than my $5k macbook pro 16

> This page was last edited on 9 May 2020, at 21:42.

Also, looking at some of the details, might this be the output of a vaporware hoax generator?

I wouldn't hold my breath waiting for this.

Ah, the EOMA68. Everyone once in a while there will be an update, but I've pretty much resigned to the fact that I'll never see the items I backed.

Sounds like EOMA68 are in the final stages of testing before they ship them out:

https://www.crowdsupply.com/eoma68/micro-desktop/updates/100...

It's only 4 1/2 years after it was fully funded, with an original shipping date six months later.

It's worth pointing out that this is a NLNet Foundation funded project; they seem to back useful projects, from the ones I've seen personally

>For all children 9-99

Reply on news.ycombinator.com