Intel unveils tiny $99 MinnowBoard Max open single-board computer | Hacker News Reader

Intel unveils tiny $99 MinnowBoard Max open single-board computer

linuxgizmos.com

209 pointsvoltagex_141 comments

This looks neat for people that want a cheap board to hack on embedded Linux. However for serious control of signal generation, acquisition, PWM, servos, etc. you really don't want to be running a multitasking OS. Something like the Beaglebone Black, with its dedicated 200mhz programmable units in addition to embedded Linux, is much more interesting for hackers and makers IMHO.

Agreed - after working with some bit modulation for a whole lot of LEDs and with the PRU unit on the BeagleBone, I don't think I'd want to use a big multitasking embedded system for finer-tuned stuff.

The PRUs are a seriously underrated feature of the BBBs. It's like having another two computers over 50x as fast as my ZX Spectrum that can run independently, and provide completely realtime output.

> The PRUs are a seriously underrated feature of the BBBs.

Speak English, please.

PRU-> programmable real time unit

BBB-> BeagleBone Black

The BBB has an extra dual core processor that runs at 200Mhz. It is interesting because it is like the processor they teach you about in your intro to computer architecture classes, every instruction is a single cycle instruction. Since it is a co-processor(not running an OS but controllable from the BBB's OS) and execution of instructions is deterministic, it is a good choice for running hard real time code.

Thanks.

Agreed - we used the PRUs for high speed data acquisition (~3.4msps.). What would have normally required a lot of outboard logic ended up being almost all software. Very much an underrated feature, although the Learning curve is not so great though for anything non-trivial.

Do you have a blog or somewhere where I can get more information on your experience using the PRUs ?

Thanks !

I'm curious what you do with them. Sage Devices doesn't seem to have a lot of info available but Bolt says y'all are doing "Home Energy Management". What in home energy management requires ~3.4msps?

I kind of wish the PRU's were Cortex-M0 or something that can be targeted with GCC. That said, they are a very nice idea.

I'm currently using the Zynq for a project and it's basically the extreme of the idea. FPGA's are kind of horrible in other ways though, such as the 30 minute build times and power inefficiency.

I considered writing a LLVM backend for the PRUs, but put it aside while I wait to see if the rumors of a TI provided compiler pan out.

Tangentially relevant, I'd love to follow more news from MirrowBoard on this project (and similar projects), but all they seem to have is a Google+ page, and I refuse to sign up for that for a handful of reasons, none which I can be talked out of.

What happened to good old RSS? I'd subscribe to that in an instant, but every time I'm directed to a Google+ feed I close the tab and forget about the company.

I'm pretty sure that for quite a few techies, looking to try out new gadgets and technology, a Google+ requirement is a deal-killer.

Use a throwaway account.

Isn't that just encouraging the problem and inflating sign up statistics?

Are you trying to protest or get the information?

That will be protest. Have some more, I've got plenty left:

Back in the good old days before the Internet was segregated into several factions behind membership walls operated by Facebook and Google we had public feeds that:

1. Didn't require signing up for anything.

2. Didn't redirect to some shitty feed aggregator that throws out partial pages so you have to click through to a page stuffed with ads and tracking to get the full content.

3. Actually respected established standards for subscriptions.

4. Actually worked properly.

Going a few years before that, we even had this thing called usenet which allowed you to subscribe to topics as well and communicate with other people posting on the feeds!!! Think of that!!!

Then Deja came along, got bought by Google, everyone got marketed to death and made to choose ecosystems, a subscriber became a dollar value or a social information provider rather than an appreciated reader, content became small and insignificant and people got lazy resulting in blindly ambling into ecosystems and getting stuck there.

And now we're creating throwaway accounts to subscribe to things because that's the only measure of value for content.

But that's not based on Federated OAuth REST Web-APIs with JSON-output and backed by a webscale cluster of eventually consistent NoSQL DBs, so obviously it's not trendy and can't have been any good.

</sarcasm>

At this point, pretty much everything about the internet is moving in the wrong direction.

That's amusing because I'm actually just fighting a "unicorn poop" proposal that puts a shit ton of sensitive financial data into the system you just described on a stack built by the lowest bidding hipsters the new TA could get in. This is to replace the 10 years investment in the existing 100% working and 5% capacity Oracle/J2EE/SOAP solution. "But but with Ruby we can rewrite this all in a couple of months" (all 2.8 million lines of quite efficient Java - of course!)

Agree with you entirely.

There were reasons why everyone started to gradually not include their whole articles in the feed, and they weren't named Google or Facebook.

I don't remember this ever really working properly, except maybe when it was called USENET.

Maybe it is time for an updated version of USENET, not tied to Google a la Google groups. I am aware that it still exists but the clients are outdated and they could use additional features.

The clients are actually all pretty good. However interest in maintaining them is not.

Eh. If you use the same browser, Google can just associate the two accounts. Also, it's more of a hassle than just saying "forget it."

It's entirely possible that if I went to a non-G+ page, I could be convinced to drop $60-100 on a whim. If you make me do stupid little things like remember credentials for a "throw-away account," then you lose the chance to capitalize on that impulse.

You shouldn't need to really remember any individual credentials for websites. Use something like KeePass.

mjolk: KeePass is a local password storage solution.

Whoops -- I was thinking lastpass. Thanks for correcting me.

One should come up with a new word for that kind of wall: privacywall ? freeregwall ? signupwall ?

Privacy pit?

Google+ has a read-only API. There are ways to generate a feed from a page. https://plus.google.com/+JaanaNystr%C3%B6m/posts/MKCV3vNsn1g

I hate that this is the top comment on something legitimately interesting. At a minimum email them, don't post here about it.

At least one good thing can come of posting here and becoming top comment: others may see it and take more care to cater to these users in the future (and since it is top comment, the commentor is apparently not alone.)

This seems like the ideal home server solution - powerful enough to run backup solutions / media server for couple people, small enough to fit wherever you want it to.

I recently went with a mini-itx BayTrail-D solution as my BTSync / Plex server - a breath of fresh air after year of fighting with ARM board in the same role. This would have been even better, as it would allow me to maintain the same form factor, instead of making way for comparatively bulky mITX case.

The only thing I'd like to know is how would you power a SATA drive attached to this?

There are external power supplies available that connect directly to a disk. I would be more skeptical regarding I/O-performance. As far as I can see, there are only two SATA-ports available.

For a cheap x86 NAS, you might want to look out for the Asrock Q1900DC which was presented at CeBit this year and is supposed to come out soon.

It has 4x2ghz Intel Celeron J1900, 4 SATA2 and 2 SATA3 ports and 19V DC input, so no more need for a bulky ATX power supply (though for a NAS you need to find a way to power those hard disks).

Judging from the prices that similarly specced boards are fetching (between 50-70€), this one should be below 100€ as well.

Here is a slide from the Asrock CeBit presentation with the specs: http://4.bp.blogspot.com/-1wOiaMWiiXw/Ux7HGzZAl6I/AAAAAAAAHQ...

I'm using an ifc6410 board as a home server, and as long as you stay with 2.5" disks all you need is 5V. The ifc and as it seems the Minnowboard as well provide 5V, so all you have to do is build your own adapter to SATA power. I just repurposed an old molex to SATA power adapter.

For 3.5" disks you also need +12V, so the easiest way would be to gut an external USB disk enclosure.

This should be a great replacement for my D510MO [1] Mini-ITX home server. It's slightly weaker in CPU speed, but it should be completely fanless and even smaller. For my needs, it should still be plenty enough.

[1] http://www.intel.com/content/www/us/en/motherboards/desktop-...

Original site is not working for me.

You can find board specs on the minnowboard official website: http://www.minnowboard.org/meet-minnowboard-max/

--update (it's working again now, slow but working) contains a lot of useful information, the link I posted contains just the specs.

Original site not working for me either. Something interesting like this gotta get a lot of traffic.

Intel News announcement (still shown at EELive in San Jose till April 3rd): http://newsroom.intel.com/community/intel_newsroom/blog/2014...

Other sites reporting the same news: http://liliputing.com/2014/03/minnowboard-max-99-intel-atom-...

http://hardware.slashdot.org/story/14/04/01/0051224/intel-up...

http://www.bit-tech.net/news/hardware/2014/04/01/minnowboard...

Nice... I'm glad we're finally getting into the 64-bit age on microcontrollers so 4+ GB of RAM isn't too far off... I've been using the latest Cubieboard (http://www.cubietruck.com/collections/frontpage/products/cub...) for a project and it's been performing pretty well, but having more RAM and x86_64 would be great.

I think we are getting 64-bit microcontrollers because everything else is 64-bit, so 64-bit has become the natural/cost-effective choice. Rather than because uC's are regularly utilizing 4+ GB. It seems unlikely that 4+ GB of RAM will be common just yet.

I wouldn't call the minnowboard a microcontroller, it's more similar to other single board computers like the Pandaboard and the odroid boards. And 2GB are already common for such boards, so 4GB are really not far off.

I would not consider the Minnowboard or Cubieboard a microcontroller. It's usually called an "application processor". Microcontrollers generally have onboard flash and ram.

32-bit microcontrollers have only recently gotten cheap enough to supplant the old 8 and 16 bit mainstays. I don't expect to see AArch64 on any of them anytime soon.

Are the chip docs open? Does linux need binary blobs on this board?

Typically the worst blob are the GPU drivers, and it seems to work with the normal open source i915 driver, which is pretty much as good as it gets.

Sorry, I missed that this had got quite a lot of attention!

http://uefidk.intel.com/content/minnowboard-uefi-firmware-eu... says

>The MinnowBoard EDKII Firmware kit is not available as a full open source release (BSD license). The CPU and chipset initialization code is currently provided as a binary module. The binary module is incorporated into the final firmware image as part of the build process.

So graphics & motherboard seem to be open, bootloader is open-ish.

As far as my EE-addled brain is concerned, there are 5100 pages of deliciousness: http://www.intel.com/content/dam/www/public/us/en/documents/...

Apparently the UEFI implementation it requires to boot is not open source :-(

Any way to replace it with a BIOS, e.g. Coreboot?

How does this compare to a Raspberry Pi?

Not sure of exact numbers but I'd guess several times faster in terms of computation speed. It runs Android which I think the RPi couldn't do successfully. USB 3 vs USB 2. Double the RAM.

And four times the price.

Three. If you want to compare this to the A model, then its four times the RAM and has wired networking as well.

From what I can see, roughly as you would expect from the price difference ie, generally higher stats.

Comparing the $99 version to the B ($35):

  - Double the clock speed (1.46GHz vs. 700MHz)
  - Double the RAM (1GB vs 512MB)
  - 8GB of on-board flash
  - Gigabit ethernet instead of 10/100
  - A USB 3.0 port
  - A SATA II port
  - A PCIe 2.0 lane
  - JTAG
  - x86 architecture

Overall, probably worth the extra cost if you need the power and features. The question is, who does? I'm considering this for no other reason than I want a board in this form factor and power class that has SATA and PCIe.

Small correction: The MinnowBoard MAX only has 8MB of on-board flash, not 8GB. (It's an SPI chip to hold the firmware)

Whoops, you are right. Reading comprehension fail on my part.

The RPI is really obscenely slow, far slower than the clock rate would suggest even for an arm. The RPI is pretty exciting as a microcontroller, though it's power usage is very high, but as a computer it's a real disappointment.

The real comparison should be with the odroid boards: http://hardkernel.com/main/products/prdt_info.php?g_code=G13... a quad arm (cortex-a9) at 1.7GHz with 2GB ram for ~$60.

Does anyone know how the performance on something like this stacks up to something like the Raspberry Pi?

A 1.4 GHz Silvermont must be many times faster than a 700 MHz ARM11.

Intel also has the Galileo board which is hardware and software pin-compatible with shields designed for the Arduino Uno* R3.

http://www.intel.com/content/www/us/en/intelligent-systems/g...

Be careful though, the Galileo emulates AVR code and is orders of magnitude slower than a real Arduino. Don't expect to pick up any shield and make it work, unfortunately.

The Galileo actually only emulates a subset of the Arduino libraries. The AVR libraries themselves are, for the most part, not supported. This makes many popular libraries unusable even when hardware is not an issue.

The Galileo's one of those boards where it's very important to pay attention to the fine print. For example, the GPIO controller is hanging off a relatively slow I2C port, so access to GPIO is much, much slower than even the lowest-end Arduino. Also, it's a modified 486 which takes multiple clock cycles to carry out many instructions that are single-cycle on modern ARM, so it's not as fast at arithmetic as the clock speed would suggest.

It has an Atom E3800 processor. How does power consumption/heating of this Atom processor compare to its competitors?

Google cache http://webcache.googleusercontent.com/search?q=cache:hLPlk9M...

IMHO what we need is not more powerful boards but more affordable ones.

At $99 it's too expensive for small projects.

maybe it still cant beat arm cpu.

arm cpu board prices are much lowerr. shenzhen cpu board is less than 30 usd.

The big advantage is x86 compatibility.

I'd pay a premium for that.

Just curious why you view that as an advantage, unless you mean to run commercial, closed-source applications on it?

Less faffing around with even open-source projects that I can't be bothered porting.

Although for me, it's Haiku that I want to run on here, and for now it's x86 :)

Even open source software is often easier on x86. There are more pre-built binaries from the Linux distributions and so on.

Because trying to run open source on almost all ARM SoCs is a massive pain in the arse. Is there a single modern ARM SoC for which you can download a Linux install image with a mainline kernel and everything `just works'?

I've had reasonable success running Arch on an ODROIDX.

http://archlinuxarm.org/platforms/armv7/samsung/odroid-x

I only tried Arch [1] on a raspberry pi so far, but there it worked just fine.

[1] http://archlinuxarm.org/platforms/armv6/raspberry-pi

Freescale i.MX. Here's their git repository:

http://git.freescale.com/git/cgit.cgi/

That's not mainline.

The almost universal lack of Mainline support is what makes ARM so frustrating. If it is not mainline, I don't want to know. Get a Minnowboard MAX and I suspect all the SoC functions will be available with a mainline kernel.

Get any ARM SoC and you are stuck with choosing between a mainline kernel with incomplete functionality OR a custom kernel that might have more functionality (if you are lucky). iMX6 might be the closest SoC to getting kernel sources for full functionality (but still stuck with blobs for Vivante GPU), but how much of that has made it to mainline (genuine question)?

If it's mainline, then you're probably not an experienced SoC developer because we've been dealing with off-track drivers and chip-level craziness for decades. We're happy to get a kernel that exposes 75% of the chip's blocks, at best.

Not sure what you mean by "full functionality" other than GPU driver source, and most of us using these for production purposes don't really care about that at this point.

Does it need to be mainline to make your product functional?

I am most definitely not an experienced SoC developer, and I am not trying to make products.

Rather my experience is merely playing around trying to use ARM-based devices as general purpose computing devices in the same way you would use an x86 device, e.g. Allwinner A10 set top box as a low power server and the Samsung ARM chromebook. In that context, I think mainline support is important. Otherwise, as someone who is not a developer, getting even simple things to work requires so much more screwing around than anything x86.

As for "full functionality", look at an SoC like the Allwinner A10/A20 and see how much of the functionality they built into the SoC can actually be used in linux, despite the excellent efforts of the Linux sunxi team.

There are open source programs out there that make assumptions as to the underlying architecture, such that they don't run as well, or at all, on non-x86 architectures, including OSes like GNU HURD. Sometimes the right tool for the job is x86.

Edit:

Grepped through the FreeBSD Ports tree for architecture specific ports, and the most common reason not related to using a binary of some sort tends to be that the program uses embedded assembly, usually floating point and/or SIMD extensions, and the developer hasn't coded a fallback of one sort or another. Some of them could probably be ported fairly easily, others may be more difficult projects where it'll be easier/mentally cheaper to just spend the extra money to get an x86 board and be done with it.

Mediocre and lousy programmers have been making implicit assumptions about architecture and unix variants for 40+ years now that cause all sorts of hell when going to other hardware. (c.f. "all the world's a vax")

Almost every package breaks on Gentoo across arches or has different stability/compatibility. Its hard stuff, the only way to fix it is to test on all these platforms which is almost impossible.

You can compile your software on a fast generic PC without setting up a cross compiler.

The fully documented GPU is a big advantage too.

What sort of project are you doing? I just picked up a teensy 3.1 ( https://www.pjrc.com/teensy/teensy31.html ) for a project I'm working on, and it's got a good amount of horsepower. If you don't need a horrible amount of power, there's also the MSP430 boards which start at ~$5, and go up from there. The low end has been covered for a decent amount of time, this board is actually interesting for the upper-end where you need something fairly beefy, but don't want to start looking at rolling your own board.

> there's also the MSP430 boards which start at ~$5

Do you have any pointers to good boards at this range?

I've used some of the Olimex ones which are barely more than breadboard breakouts with a good amount of success.

Damn, it looks like they have increased the price on the launchpad since I picked it up. It's now $10. It comes with two micros though with separate use cases, so you can use the launchpad for prototyping before putting it on a breadboard.

This could be excellent as Plex client if it does support HDMI CEC. Raspbery Pi is too slow for this

Okay, this is perfect for my low-cost Haiku box project. I'm definitely getting one to hack on. Well done, Intel, and the whole "launching with Android and Linux" thing seems to be indicating an interesting shift...

Does anyone know what would be a reasonable network throughput to expect on this device? I see it has a 1Gbps port, but whether or not the CPU can saturate that is not obvious.

Interesting to see it announced with Linux and Android support.

If this doesn't signal a shift from a Microsoft-oriented (Intel) world to a new era, I don't know what does.

I think people are trying to read too much into this device. This is one of many small board computers you can buy today, many of which particularly in the industrial market have x86 processors. If this sells a few hundred thousand boards like the Beaglebone Black it will be a smashing success. This isn't a statement about Intel moving away from MS.

I think it signals diversification for Intel; not necessarily a shift away from Windows. Running Windows on ultra-low-end boards like this makes no sense; the license would cost more than the board itself and the board isn't powerful enough to run anything approximating a modern version of Windows. It's an order of magnitude slower than your cell phone.

Ultimately though, the x86 PC (aka Windows) market is not a significant growth market for Intel. They'll keep pumping money into it as long as they get their 3-5% growth, but if Intel wants to move the needle for the company, they need to find new markets.

> It's an order of magnitude slower than your cell phone.

Really? I'd like to see some evidence of that. I'd assumed they were roughly the same ballpark.

The base $99 model is a single core 1.4 gHz E3815. The graphics core is significantly slower than a cell phone SoC as well. Intel is not an inferior company or anything; just this product at this price point is slower than a 3 year old cell phone.

I do wonder what would it take to run Windows (oh perhaps XP) on Minnowboard.

The drivers would be the biggest issue. Licence cost would not be that bad(MSFT was charging $15 per licence to netbook vendors a while back).

This shift happened a long time ago.

Intel has been by far the best supporter of Linux of any SoC / processor vendor, with first party support of open source drivers for their wifi and GPUs. They patch the kernel to support their devices far ahead of their actual release. They have been pushing their own Linux based OSes (Meego, Tizen) for a while (without success).

I'd love to see a small, low power board (ARM or Intel) with Dual SATA ports, ethernet and preferably a serial console instead of the HDMI port.

Dual SATA seems to be the sticking point. I'm still using a GlobalScale Dreamplug [1], but you'd be insane to pay retail (~250US) for it now. It's an 800mhz ARMv5 with 512mb of RAM and a single SATA port. It's well supported by the mainline kernel (with an upgrade to uBoot).

The FreeScale iMX6 would seem to have enough I/O for two SATA ports (apparently it even has PCIe!) but I've never seen this in a retail board.

At this stage you might be better off going for an off the shelf NAS, something that can run FreeBSD and ZFS.

[1]: http://www.globalscaletechnologies.com/t-dreamplugdetails.as...

HD4000...hmm can we play Diablo III on it?

Most likely. Skyrim was playable on the HD3000s, in my experience, so...

As I said above, this thing is much slower than the HD 3000 or 4000.

I think memory might be your bottleneck with newer games. Apparently Diablo III will tolerate having just 1GB, but only on XP.

It's not HD Graphics 4000. It's just HD Graphics, with no numbering. It only has 6 execution units (compared to 16 on HD 4000) and is clocked much lower.

Actually, per the datasheet[1] it has 4 EUs. I could not find any information about clocking the core, but I didn't search the entire 5100(!) page document.

[1] http://www.intel.com/content/dam/www/public/us/en/documents/...

I thought it was the same stuff they put in the Celerons and Pentiums, just clocked lower, but if only has 4 EUs, that's even worse.

I can see this being big with the media-art crowd for installations if it has any sort of real graphics oomph. From what I've heard the pi is just barely capable of hd video, anyone able to divine these tech specs into a comparison with the pi's graphics capability?

Hmmm... can it run Windows 7? Serious question, I'm after a cheap base system for a HTPC and this looks quite like a fit.

If you built a UEFI stick (basically drop the contents of a Windows 7 DVD onto a FAT32 (not NTFS) drive) then maybe. XBMC will run on this, but I'm not sure if it can decode 1080p H264.

Is it PC-compatible enough to run DOS? If not, then probably Windows is entirely out of the question. Just because it's x86 doesn't mean it's PC-compatible.

can you connect a bunch of these in parallel and run, like, map-reduce or something ?

Sure, that's called microservers. It's not a good idea, but you can do it.

Why aren't microservers a good idea?

Why is it not a good idea? Genuinely curious

If I had to guess, probably because you could get equivalent performance cheaper with standard servers. Plus, there's a good chance that your distributed solution could fit on a single server, removing the overhead.

Yep. If you compare a $1,000 Xeon E3 server with a $1,000 cluster of ten Minnowboards, the E3 will deliver more performance. Likewise, a $10,000 E5 server beats a cluster of E3s in many cases.

what sort of devices it would support mobile or laptop?

Neither, its really designed for general purpose hacking, connecting to external electronics, sensors, motors, etc. just like the Raspberry Pi ...

Is there any way of upgrading the RAM on this so that it can run SteamOS?

I don't think an Atom SoC would be adequate in terms of processing power. This is for more modest chores.

I had hoped to use BeagleBoneBlack for a mobile low-powered application, but the overhead of the different architecture was too much hassle, and I ended up using an Intel NUC.

With NUC I can use the same OS as our high-end servers, same 64-bit apt-get repos, and Docker.

MinnowBoard looks like what I've been waiting for - fingers crossed...

Well the Pi had a good run, honestly longer than I thought it would but the big boys have entered the arena now.

Also imagine what the prices on services like kimsufi and online.net will look like once these guys have hundreds of thousands of them rack mounted. Good times to come.

The PI is still a third of the price, depending on you want to do with it, it might be amply enough.

For how long though...

There's also the ODROID-U3, which is roughly comparable to the Minnowboard, and it's just 59$.

http://hardkernel.com/main/products/prdt_info.php

But do note that it is a "community price, one for customer" only offer. I doubt they are losing money on it, but I suspect you'd have to buy a few hundreds (or maybe thousands) to get that price if you actually want to build something around it.

The regular price is $65.

http://www.hardkernel.com/main/products/prdt_info.php?g_code...

Maybe, don't forget that the objective of the Raspberry Pi Foundation has always been education, and keeping the same hardware platform for a longer period of time, eschewing participation in a hardware arms race, works well for this objective (in terms of building learning materials for teachers around R-Pi etc).

That said I think they have said that there will be an R-Pi too sooner it later, 2015 perhaps?

This is a great point, R-Pi foundation may not even need to make it's own boards. This could very well work out best for everybody.

The Pi has advantages that are not easily overcome by this:

Price: $25 for a Model A. Yes, it delivers significantly less - but at $25, it's a consumable for projects, whereas at $100 the Minnow probably isn't. $35 iff you want Network and another USB port and some more memory.

Community: If you want to do anything with the Pi, you'll likely found someone who has documented doing something very similar, and probably uploaded their source code.

Peripherals: There are a lot of them for the Pi, relatively cheap.

I think the Minnow Max will carve its own (smaller, but still nontrivial) niche, but will not actually harm RPi's market share.

Community is _extremely_ important; Arduino lives on its community despite losing on pure technical benchmarks against all sorts of other dev boards.

Pi isn't going anywhere, as they have a great understanding of the community they wish to create and do educational projects with.

I can recommend the Pi to someone who has never used Linux before because there are so many resources available that are Pi-specific. It's the same reason I recommend Ubuntu for newbies even though I prefer other distros for myself.

I started writing a RoR medical office application. Then I thought what if I wanted every patient to have an RFID card and perform measurements (BP, Weight, etc.).

I found an RPi + Arduino project with peripherals, manuals and all I could ask. That I can handle, building everything from scratch (from software to hardware) would take a tremendous amount of time (due to lack of expertise on my part).

So as you said, even if you get me an x86 board at 50 USD I'd still go for Arduino.

Not to mention Arduino seems to have shown commitment to their pinout configuration. I seriously doubt later iterations of this board will retain compatibility.

This board costs 3 times the RPi (model B), it's not comparable. It's x86 so the premium price is justified IMHO, BUT it's not really comparable with a 25 USD board.

Maybe true for a personal user, but not for a business.

If you're not basing a product around it, the parts cost is trivial compared to the engineering time cost. If you want just one (or a handful) for a one-off project, $25 and $75 are "the same" to any sane business for this. The team will likely spend more on coffee than controllers/SBCs in the course of development.

I see you have never worked for the accounting department at certain large companies like I used to work for.

Oh, we had to buy our own coffee, too.

While I agree with you, my experience is that most businesses don't recognize your definition of sane. My office decided that providing ink pens was too high of an expense.

The Pi uses 2 watts [1], this uses 5. There are a lot of uses where that 3 watts can be rather significant. They both have niches, and it's going to be interesting to see which one this carves out.

[1] http://www.raspberrypi.org/phpBB3/viewtopic.php?f=63&t=6050

The "big boys" of embedded SoCs (Samsung, Atmel, NXP, Freescale) have been around a lot longer than RPi and are doing just fine.

Reply on news.ycombinator.com