Slashdot Mirror


New $149 NetBSD Single Board Computer Port

An Anonymous NetBSD User writes "NetBSD now supports a new ARM board, the Technologic Systems TS-7200. The TS-7200 is a low cost ($149!) mass produced embedded single board computer that runs on less than 2 watts of power."

113 comments

  1. Fun! by BrynM · · Score: 1

    Makes me want to build a new firewall/ssh server. Now what interesting thing to mount it in...

    --
    US Democracy:The best person for the job (among These pre-selected choices...)
    1. Re:Fun! by Anonymous Coward · · Score: 1, Informative

      Won't be so fun to have a firewall with a single NIC.

      Might as well get a Soekris net4801 (www.soekris.com). You get two more ethernet ports (three total), a faster processor (266Mhz x86 compat vs. 200Mhz MIPS), more RAM (128MB vs. 32MB), and a compact PCI and a real PCI slot (3.3v cards only).

      It costs $50 more, and you get tons more features, most of which are mandatory for a firewall.

    2. Re:Fun! by BrynM · · Score: 1

      Good tip! Thanks!

      --
      US Democracy:The best person for the job (among These pre-selected choices...)
    3. Re:Fun! by Anonymous Coward · · Score: 0

      Won't be so fun to have a firewall with a single NIC

      I'd say it's probably the most secure firewall you can get.

  2. Direct Link to the Board by Anonymous Coward · · Score: 5, Informative

    Is Here

    Keep It Real, Lameness Filter(TM)

    1. Re:Direct Link to the Board by Bastian · · Score: 2, Insightful

      It's not expensive, but $150 doesn't seem like Rollback Prices for what you get. A nice microcontroller or all-in-one embedded computer.

      Granted, you have to pay quite a bit extra ($180) to get the kit with a power supply (gotta have one of those), and a CompactFlash card (also gotta have one of those. 8MB RAM(or 16 if you get the $165 version) is going to fill up quickly.)

    2. Re:Direct Link to the Board by Anonymous Coward · · Score: 1, Informative

      Its 8MB of on-board *FLASH*, not RAM. There is 32MB of RAM. This flash is always on the board regardless of whether the compact flash is populated. There is an RTOS (eCos/RedBoot) on the onboard flash and you can still boot the board via NFS root even without anything in the CF.

      How many microcontrollers do you know of that will run Apache?

      5V power supplies are ubiquitous and can be purchased for like $4 at allelectronics.com. CF cards also and can be purchased for $26 elsewhere, even Target has them. There is nothing about the board that would imply you have to buy the accessories too from Technologic Systems.

  3. Re:Direct Link to the Board and specs by Isosonys · · Score: 2, Informative

    TS-7200 Single Board Computer 200 MHz ARM processor for Linux ARM9 processor with MMU 32 MB SDRAM 8 MB flash drive (16 MB optional) 10/100 Ethernet Compact Flash 2 USB host ports 2 COM ports 20 DIO PC/104 expansion bus Optional A/D and RS-485 Optional 802.11b WiFi

  4. Funny this should come up by skinfitz · · Score: 3, Funny


    I've recently been looking at small / quiet form factor boards from places like Mini-ITX - I'm embedding an X terminal into my glass topped dining room table.

    I've had it with desktops; time for the X table top.

    1. Re:Funny this should come up by Anonymous Coward · · Score: 0

      Why not embed your X terminal in a glass wall and call it an X window? *groan*

    2. Re:Funny this should come up by forkazoo · · Score: 1

      That sounds like quite an interesting project. I have been pondering for some time building a complete media center setup, with an HTPC plugged into the TV, my main server with all my movies and saved shows, and a small PC embedded in the living room table to act as the mother of all remote controls, so even my dad could use it because everything would be handled in one place, from changing the channel, to browsing archived videos on the server, without ever being able to lose the remote, or carry it into the kicthen and forget it in the freezer, etc. ::sigh:: I'll need to take about 20 years off work, and hire a team of helpers to take care of all the projects like that I want to do. BTW, your comment sounded like an interesting enough project that I made you a friend. Good Luck!

  5. Crack-influenced flash prices... by rossifer · · Score: 5, Informative

    The computer itself seems like a steal for the specs they're claiming but their prices on CF flash are more than a little silly. You can get Sandisk Ultra II 512MB for $60 from any online meta-retailer and these guys are offering an unknown 512MB CF card for $145.

    I've seen this kind of thing from several of these kinds of suppliers and I don't quite get it. They don't seem to realize that things like flash are fully commoditized and are still thinking they can get away with a 100% markup...

    Oh well, if they can actually sell CF at those prices, then they deserve the money. Kudos for such a sweet, low power, inexpensive computer either way.

    Regards,
    Ross

    1. Re:Crack-influenced flash prices... by skinfitz · · Score: 2, Interesting

      I was thinking the same thing - reminds me of Apple RAM pricing. Obviously the thing to do is buy it without the CF if possible and get the CF cheap elsewhere.

    2. Re:Crack-influenced flash prices... by Anonymous Coward · · Score: 5, Informative

      I work at embeddedARM.com, we know we're aren't competetive on CF prices and even suggest to cost sensitive customers to go get their CF's from Costco or what-not. We sell them at all just so we can be the one-stop shop to customers if they need them with the board, and pre-installed with Linux, etc..

      Also, some (not all) of our CF are inflated because they are the the "industrial" class CF's. These have extended rewrite cycles, higher G ratings, and work in extended temperature.

    3. Re:Crack-influenced flash prices... by eviltypeguy · · Score: 1

      Reminds you of Apple RAM Pricing?

      Take a look at Dell's RAM pricing sometime.

      My company always buys our Dell machines with the minimal amount of memory and then buys the amount of memory we're actually going to use from somewhere local.

      I'll never forget one time when we saved almost $500 doing things that way on a single server...

    4. Re:Crack-influenced flash prices... by Anonymous Coward · · Score: 1, Informative

      The industrial grade CF's are from San Disk

      http://www.sandisk.com/industrial/flash-drive.as p

    5. Re:Crack-influenced flash prices... by TheRaven64 · · Score: 1

      Most system manufacturers charge a lot for RAM. If you want real extortion, take a look at Sun video cards. Recently they were charging around $400 for a Radeon 7000, which is identical to the Mac version of the same card (retailing at around a fifth of the price).

      --
      I am TheRaven on Soylent News
    6. Re:Crack-influenced flash prices... by ffsnjb · · Score: 1

      HP is just as bad. I refuse to ever power on a new machine without the max amount of RAM the board will take. When I bought my new laptop, HP wanted $1600 for 2 GB of RAM. Crucial wanted $998 for their awesome RAM. HP shipped me the laptop with a 256 MB stick I replaced with the Crucial before powering it on. I saved a great deal of money and have better RAM in my machine.

      Now, I just need to take the time to ebay that 256 MB stick.

      --
      "Why do you consent to live in ignorance and fear?" - Bad Religion
    7. Re:Crack-influenced flash prices... by Danta · · Score: 1
      PC Engines offer single board computers from $120 and up.

      Their 2-NIC model (for use as firewall) goes for $140.

      Their flash is not overpriced either ($17.25 for 128MB CF).

      I run OpenBSD on one of those, so I am pretty sure NetBSD should run on them too.

    8. Re:Crack-influenced flash prices... by innosent · · Score: 1

      Anyone know of similar products for Gigabit Ethernet?

      --
      --That's the point of being root, you can do anything you want, even if it's stupid.
    9. Re:Crack-influenced flash prices... by Anonymous Coward · · Score: 0

      Yeah, I mean, no other companies charge a preminum for additional RAM. That's not industry standard at all. Oh wait...

  6. We can run Debian on it already by at2000 · · Score: 1

    it does give us more choice, but not too significant.

    1. Re:We can run Debian on it already by LizardKing · · Score: 1

      I'd be very suprised if Debian actually supported it as well as NetBSD, based on the dubious support for non-i386 architectures. Debian's Sparc support is hit and miss, and their SGI MIPS support is a joke.

    2. Re:We can run Debian on it already by Anonymous Coward · · Score: 0

      Which brings up the really key point:

      Support != "It compiles binaries for the CPU!!!11!11!!!eleven"

      If you ain't got the drivers, it's not going to work.

    3. Re:We can run Debian on it already by LizardKing · · Score: 5, Insightful

      Which is why you should go with Linux, rather than NetBSD, since NetBSD's notion of "support" runs to "we can get it to boot on this hardware" rather than "works well on this hardware and supports all common drivers"

      Your description of NetBSD's hardware support is a bit odd. Unlike Linux where the philosophy is "if it works on one platform include it", NetBSD is designed so that a driver written for one platform will work reliably on all those that support the same device. As for missing support for some devices, that's no more or less true than Linux. Compare the support for SGI MIPS for example - when it works in Debian it supports a similar subset of systems and devices. Likewise for the Vax, although here Linux supports a tiny suset of the systems and devices that NetBSD does. Linux has a large proportion of platforms that are only notionally supported (they booted once upon a time, but have languished thanks to Linus being primarily concerned with i386).

      all things considered, Gentoo is a better bet

      I don't know of a single embedded company using Gentoo (embedded boards being the subject of the original article), and all the embedded Linux kits I've used have been based on RedHat. I also know of no companies using Gentoo on their servers - again it is RedHat along with the odd SuSE box. I did encounter Debian once as a contractor, but the guy who had installed it was reputedly a loose cannon and ahad done it without auhorisation (one of my first tasks was to switch the machine in question to RedHat).

    4. Re:We can run Debian on it already by at2000 · · Score: 1

      That depends on how you look at it. In Debian, packages are usually set to support all arches, unless they are known not to work. In Gentoo, many packages are not available for some arch. It is difficult to compare which is better: "all supported packages in Gentoo works" or "some packages in Debian works".

      If something does not work, you can always fire a bug, whichever distro you are using, but if the platform is more "standardized", this happens less frequently.

    5. Re:We can run Debian on it already by Eneff · · Score: 1

      It looks like Embedded x86 is behind this kit. I have a suspicion that they were behind the porting effort, and that would mean that they made damn sure that NetBSD works flawlessly on the thing.

    6. Re:We can run Debian on it already by Anonymous Coward · · Score: 5, Informative

      (I work for Technologic Systems)

      I did both the Linux port and the NetBSD port to the TS-7200 so I am uniquely qualified on their differences. IMHO, to an engineer actually doing the work, NetBSD is a far superior OS to port to, at least to this platform.

      The port is not a "we can get it to boot" port of NetBSD. There is installatium medium on a FTP site, a system installation program (sysinst), cross-toolchain support and even support for the ISA (PC/104) bus, which is something that is really impossible right now on Linux due to its x86 assumptions littered throughout ISA drivers. (Ever tried to use inb and outb on an ARM?, there is no such thing! ISA bus space is memory mapped and appears at a different spot depending on 8 or 16bit accesses)

      Although you can run it on Debian, Debian does not support it. I had to do all that work myself, and Technologic Systems is supporting it on this platform. There aren't any TS-7200 kernels on debian.org.

    7. Re:We can run Debian on it already by Anonymous Coward · · Score: 0

      As a UC Berkeley graduate in Philosophy, you are quite incorrect. Let's schedule a debate sometime about which operating system is the better to port. All matters are open to debate.

    8. Re:We can run Debian on it already by Anonymous Coward · · Score: 0

      When the platform is "standardised" to the point that core toolchain component versions are selected based upon what works best for x86, this happens a hell of a lot.

    9. Re:We can run Debian on it already by Anonymous Coward · · Score: 0

      Is the Linux port based on TS-Linux? If so, has it been cleaned up or does the customer still have to correct all of the errors, like the hwclock typo in rc.sysinit, that have been there since at least 2.0?

      Will TechnoLogic be working on NetBSD for their x86 line?

    10. Re:We can run Debian on it already by Anonymous Coward · · Score: 0

      Dude. How many drivers do you need for one embedded platform? And why can't you fix them up for your platform considering they are already there and working for i386?

      If you can't do that then you must be a sad, sad man.

      I had to do all that work myself, and Technologic Systems is supporting it on this platform. There aren't any TS-7200 kernels on debian.org.

      This is your problem. Why don't you submit your fixes back to the kernel tree? Nobody is going to suddenly make all your ISA drivers work on your platform out of the goodness of their heart. Listen: nobody cares about doing you a favour, especially if you can't do one for anyone else.

  7. SBCs by dmiller · · Score: 4, Informative

    This looks like a nice SBC, but I am really happy with my little Soekris'. With case, the 4501 is cheaper than this ARM board (the board alone is more expensive) and has three ethernet interfaces.

    1. Re:SBCs by batobin · · Score: 3, Informative

      I have to agree. All the networking guys I know love these things.

      I've even talked to one guy who was using one at a remote radio tower which serviced an entire town of wireless users. He had a cheap solar panel hooked up to a little battery, both of which powered the soekris for years without problems.

    2. Re:SBCs by aivarsk · · Score: 3, Informative

      Except that Soekris 4501 uses 5 times as much power as TS-7200. o.O

    3. Re:SBCs by Anonymous Coward · · Score: 4, Informative

      (I work at Technologic Systems)

      We are a company purely of computer engineers and cheap case design is not really our specialty. We are indeed investigating cheaper case designs and 2/3 ethernet variants and in the future will have no trouble undercutting Soekris. :-) This is somewhat of a new market for us as we have traditionally targetted industrial embedded uses (hence PC/104 and RS485 com ports), where 3 ethernets is a bit of an odd request.

    4. Re:SBCs by Anonymous Coward · · Score: 0

      The case is also more expensive because it is not just a case, but includes a signal conditioning power supply with industrial style connectors.

    5. Re:SBCs by justins · · Score: 1
      He had a cheap solar panel hooked up to a little battery

      How much does a setup like that cost? Where can I find out more? Just curious, there's a lot of useless crap when I try to figure it out with google. It'd be great to hear from someone who has been successful with it.
      --
      Now before I get modded down, I be to remind whoever might read this that what I am saying is FACT. - bogaboga
    6. Re:SBCs by Anonymous Coward · · Score: 1, Interesting

      solar generally goes for no less than $3.50/watt when you buy the big (150-200W) panels, and more when you're getting a smaller one... you'll need to multiple the board power requirements by six or eight to account for night, clouds, etc.
      then you just need a simple lead-acid car battery, anywhere from $50 to $150 IIRC. and a charge controller to keep the panels from overcharging the battery.

    7. Re:SBCs by Anonymous Coward · · Score: 0

      the Geode in the Soekris is in a whole different class than the ARM toy... 15W vs 1.5W. for 15W you might as well just get an old 400MHz laptop, it comes with PCMCIA...

      also a 15W board might require 120W of solar depending on where you are, which is a mite more expensive than if you were using an ARM.

    8. Re:SBCs by dmiller · · Score: 1

      Cool - I'll watch with interest!

    9. Re:SBCs by Anonymous Coward · · Score: 1, Interesting

      I have a couple of freebsd router appliances I have built with 4 interfaces (separate nics for internet, my trusted subnet, guest subnet, and one to wireless bridge). It takes a substantial amount of CPU to route and firewall 100Mbs and more for each interface and especially if the appliance generates the traffic via samba or nfs. Also, not all nics are equal, I've seen 3x speed difference between different nics, probably dependent upon hardware checksumming.

      I'd love to replace my appliance with a low power version, but in my experience 200MHz won't nearly cut it for file serving plus serious routing.

      I'm probably not quite the customer for this product, but I would (will...) buy a moderately priced, open, low power, 4 nic appliance with good disk & network I/O. Till then I'll draw 100 watts+ on surplus 500MHz box, yuck!

    10. Re:SBCs by stratjakt · · Score: 2, Insightful

      3 ethernets is an odd request?

      Isn't the "Internet, LAN, and DMZ" paradigm pretty common for routers/gateways?

      --
      I don't need no instructions to know how to rock!!!!
    11. Re:SBCs by mph · · Score: 1
      3 ethernets is an odd request?

      Isn't the "Internet, LAN, and DMZ" paradigm pretty common for routers/gateways?

      Try reading the whole sentence. He isn't talking about the router/gateway market.
    12. Re:SBCs by MrResistor · · Score: 1

      Isn't the "Internet, LAN, and DMZ" paradigm pretty common for routers/gateways?

      For routers/gateways, yes. For something like, say, an irrigation system controller, even one ethernet is somewhat unusual.

      --
      Under capitalism man exploits man. Under communism it's the other way around.
    13. Re:SBCs by leitz · · Score: 1

      Can you put a guess time on "in the future". This interests me but putting things together might get expensive quickly. Being new to the idea I'm not sure where the inexpensive routes are.

    14. Re:SBCs by funky+womble · · Score: 1

      When it says "max 15W" on the Soekris product page for the 4801, that's the maximum (I think it's the power available from the DC-DC converter), not a typical figure. They're more like 5W in use (more with a HD of course). There are PCMCIA Soekris boxes too, but if it's wireless you want it for, MiniPCI is often a better idea.

  8. Actually, 1.875 Watts. by Perdo · · Score: 2, Informative

    Comparable to VIA's Eden-N at 2.5w at half the size.

    I'm pretty sure NetBSD has already got an x86 port too...

    An extra $50 can buy a lot more technology elsewhere.

    --

    If voting were effective, it would be illegal by now.

    1. Re:Actually, 1.875 Watts. by justins · · Score: 1

      Yeah, but let's be fair. The Eden-N is meant for end-user applications, for the most part. It's built in such a high volume that it should be pretty cheap.

      The TS-7200 targets a different market. I don't think anyone is under any sort of illusion that this is going to sell a lot of units; it's a prototype board for developers. If you judge it by the standard of oddball-architecture boards, its price is extremely reasonable.

      Doesn't seem very good for development, though, with only 32MB of RAM.

      --
      Now before I get modded down, I be to remind whoever might read this that what I am saying is FACT. - bogaboga
  9. Linksys WRT54G by Anonymous Coward · · Score: 0

    is this quite similar to the hardware that runs in the Linksys WRT54G linux based router?

    Would be nice to be able to run netbsd on it.

  10. True, but have you checked out the competition? by epseps · · Score: 3, Interesting

    Here are some ARM based boards to compare:

    Microcontrollershop.com

    Here is a similar Atmel board but this is only 8 bit and $153. Atmel Ethernut Board

    8MB or 16MB flash with 32MB ram on a 32 bit processor is a good deal in the world of small low powr computers. (but still expensive compared to x86 desktop)

  11. Embedded NetBSD by Scott7477 · · Score: 1

    Good story; this is what I like to see from Slashdot is links to interesting products.

    --
    "Lack of technical competence coupled with the arrogance of power, as usual, leads to no good end."
  12. Lower priced board also available. by Sonic+McTails · · Score: 2, Informative

    If you want to go cheaper, got fo an NSLU from Linksys (http://nslu2-linuxorg). It's a 100 dollar ARM board w/ Ethernet, and two USB ports. It comes running Linux, but there has been some work with getting NetBSD running on it according to the mailing lists. It's a 233MHz ARM, 32MB RAM, 8MB Flash. It also has 1 1/2 serial port which you can solder on with little effort. The 1/2 serial port is only pinned out for an RX, no way to transmit with it, so it's kinda useless in that regard.

    --
    This signature was left intentionally blank.
    1. Re:Lower priced board also available. by Anonymous Coward · · Score: 0

      There are several problems with using the NSLU in a real embedded application.
      1) it is consumer electronics and may stop being manufactured or undergo major internal changes at any time.
      3) No compactflash socket
      4) it is consumer electronics
      5) No PC/104 expansion. Almost all real embedded applications need some sort of way to add custom hardware on a FPGA or whatnot.
      6) No way to do DIO or ADC.
      7) it is consumer electronics

      The NSLU2 is not even the same industry as the TS-7200. I wouldn't want to put one of these anyplace other than at home to play with. Even then, if I really wanted to play with something in this category, I'd probably go buy an XBOX to run Linux on as this is even more impressive than the NSLU2 and the TS-7200. I hear Microsoft is going to drop the XBOX to $99 sometime this year.

  13. What?!? by Anonymous Coward · · Score: 0

    I have to wonder what basis you have in assuming Technologic will "make damn sure that NetBSD works flawlessly".

    I _hope_ they do a good job but my experience with TS-Linux tells me not to be surprised if they don't.

  14. Temperature range by codemachine · · Score: 1

    This SBC looks really great for a project I might be taking up, but there is one problem.

    I need it to be able to withstand operating temperatures as low as -40C, since it will be outside in the winter (and later this week, it will be -38C here, so it really does need to handle -40C).

    I suppose some sort of heater in the enclosure would be one way to deal with it, but I'd love to know if this board or a similar one could handle the temps without any extra effort.

    1. Re:Temperature range by norkakn · · Score: 1

      is power a huge concern or could you put a lightbulb in the enclosure?

    2. Re:Temperature range by codemachine · · Score: 1

      Power isn't too big of a concern. I'm just wanting to keep costs down.

    3. Re:Temperature range by Anonymous Coward · · Score: 0

      (I work for Technologic Systems)

      We do sell extended temperature versions of the board, but you have to buy in quantity. We've found in practice the boards will work to the extreme low temperatures without even using extended temp parts, its usually the extreme high temps they are more likely to have a problem with.

      We've had a customer also put a big ceramic 25watt resistor across the 5V power supply to act as a heater which also works just fine.

    4. Re:Temperature range by Anonymous Coward · · Score: 0

      As AC i doubt you'll ever read this.. but if you look at the site under options theres an "extended temperature range" -40C -> 85C for $30

      sounds like what you are looking for.

    5. Re:Temperature range by codemachine · · Score: 1

      Thank you all for the replies.

    6. Re:Temperature range by Anonymous Coward · · Score: 0

      But do the parts keep working with condensation? The intended low-temp application is also an outdoor application. Temperature and humidity changes would be what would kill the system. It's not your area, but it'd be nice if someone could point folks in the right direction. (I can't. Never built something like that, just heard horror stories about nesting chipmonks.)

  15. Openbsd on the 2nic: by holysin · · Score: 1

    I have to ask, how much traffic are you firewalling with the 266 chip?

    and of course, given much thought into selling your own brand of firewalls? ;-)

    1. Re:Openbsd on the 2nic: by Danta · · Score: 1

      I just use it for my 512Kbps ADSL line, don't know how much traffic it could handle max.

  16. Time to revise these stories by niteice · · Score: 1

    It's getting to the point where we shouldn't hear about the architectures NetBSD supports, but rather, the ones it doesn't. ;)

    --
    ROMANES EUNT DOMUS
    1. Re:Time to revise these stories by setagllib · · Score: 2, Informative

      This is a single board, not really an arch...

      NetBSD not supporting an architecture means one of two things: (1) the architecture has no processors with an MMU, or (2) no developer has the hardware

      This explains no NetBSD on iPaq/etc for (1) and no NetBSD on ppc64/ia64 for (2).

      As we can see by these examples (especially that nice post by the developer who supported this board under NetBSD), when it does support the hardware, it supports it properly, not just as x86 compatibility. That's the key difference between Linux and NetBSD; NetBSD is an operating system for every architecture, Linux is a kernel that can pretend to understand any architecture. I'll stick to NetBSD, thanks (even on i386 where possible).

      --
      Sam ty sig.
    2. Re:Time to revise these stories by Anonymous Coward · · Score: 0

      Hey bozo, NetBSD isn't the be all and end all that you would like to have us think.

      There are plenty of areas where NetBSD is behind Linux in portability and hardware support and performance. Like this simple test where Linux quadruples its performance.

    3. Re:Time to revise these stories by setagllib · · Score: 1

      Later in that same thread: http://marc.theaimsgroup.com/?l=netbsd-tech-kern&m =110618215623555&w=2

      Should be more careful about what you post. Doesn't help your credibility, Anonymous Coward.

      --
      Sam ty sig.
    4. Re:Time to revise these stories by Anonymous Coward · · Score: 0

      What credibility you may have once had is now non existant. So shut your crap holes.

      As of now, nobody ever has to prove anything to you again. Anything you say is to be assumed wrong until you construct an air tight proof for it. But I have a better suggestion: shut the hell up. Really. Just stop posting.

      Have a nice day.

  17. That's just the cpu by phr1 · · Score: 1

    Looks like the Via motherboards use around 10 watts watts because of all the heavy video stuff on them (they are intended for media centers etc). On the other hand the ARM chip has no floating point processor. I've been looking for a small, low-powered board that can run an Ogg Vorbis encoder, which uses lots of floating point. Nothing looks attractive yet.

    1. Re:That's just the cpu by Perdo · · Score: 1

      Pentium M:

      http://www.lippert-at.com/index.php?id=21

      Best FPU per Clock of any current microarcitecture:

      http://www.cpuid.com/PentiumM/index.php

      At 2 Ghz it should be to OGG encoding what a GPU is to Graphics Rendering.

      - expect to pay laptop prices for it -

      How fast do you want to spend?

      --

      If voting were effective, it would be illegal by now.

    2. Re:That's just the cpu by phr1 · · Score: 1

      I just need to encode Ogg in real time. My P3-750 laptop does it at maybe 5x real time, so I need the equivalent of about a p3-150. A 400 mhz XScale would probably be fine if it had floating point. Know of anything?

    3. Re:That's just the cpu by setagllib · · Score: 1

      You might get away with a faster processor but using software floating point instructions. This way you can stick with a clean architecture and still get the floating, just with less efficiency.

      But if it was up to me I'd look for a MIPS R5000. They have good floating point (powered the multimedia machines of the last generation, mostly SGI boxes) and maintain the advantages of low-power low-heat processors. They can also run in 64 bit mode but not many free OSs support this properly.

      Try this out: Find yourself an SGI Indy with an R5000, stick NetBSD 2 on it (or Linux, but you're stuck with 2.4, and nothing is maintained any more), and use that. Indy machines have very good sound cards you can use to play the OGGs too... and the only fan part is the PSU. It SHOULD be powerful enough (at 150Mhz, the R5000 has integer ops *roughly*equivalent* to a P1 166, but floating ops are much better and without the bugs).

      --
      Sam ty sig.
    4. Re:That's just the cpu by Anonymous Coward · · Score: 0

      I'm sorry, the SGI Indys with R5000 processors are supported by Linux 2.6 and continue to be maintained.

      Just because you are too incompetent to make it work doesn't mean it isn't supported.

    5. Re:That's just the cpu by setagllib · · Score: 1

      http://www.linux-mips.org/archives/linux-mips/2004 -09/msg00112.html
      That's from a friend of mine off another forum.
      Gentoo's documentation on running MIPS agrees that 2.6 isn't flying on it. In 2.6.10 it might be fixed, but the changelogs don't show anything like that - after all, it's old and irrelevant right? Who cares about hardware not actively in use by big corporations?

      --
      Sam ty sig.
    6. Re:That's just the cpu by Anonymous Coward · · Score: 0

      Yeah, why don't you tell me about this weird new invention the i386 has called "page tables"? The MIPS port must not be emulating these so called "page tables" correctly or something.

    7. Re:That's just the cpu by Anonymous Coward · · Score: 0

      http://www.linux-mips.org/archives/linux-mips/2004 -09/msg00112.html

      Wha?? You show me a single example to make your case? I can give you an example of an i386 system that doesn't boot on FreeBSD or NetBSD. So does that mean neither OS supports the i386 platform properly?

      (And yes, I can find examples where they used to work and no longer do so. It isn't that uncommon if you trawl through mailing lists.)

      No. What it means is there is some small bug or oversight in the platform bootup code, or some slightly different machine that developers don't have access to or documentation for. Same deal with Linux and MIPS.

      There are plenty of people running 2.6 on MIPS, so you can't pull out one counter example to make your entire case for you.

  18. Pay the extra $30 by Anonymous Coward · · Score: 0

    For $30 on their site it says you can get an "extended temp range" which is -40 to +85 C. That's -40 to +185 for you non-metric people.

  19. Re:SBCs & Solar Power by drwho · · Score: 1

    Designing a good solar power system is not a trivial task. Obviously you need panels, but there are polycrystalline and mono, the mono being nicer but more expensive. There's the weatherproofing: you need a strong covering that is quite transparent over a wide colour range - even UV, and something that will not decay over time, nor allow water to bead. The support mechanism needs to be strong, and provide maintenance access. Photovoltaic cells are much more efficient when cold, so ventilation is important, but with ventilation comes weatherproofing woes.

    Next problem is that the sun moves across the sky, and for maximum efficiency you want to track it, which can be done in one of several ways. Many people use a motor to tilt the panel, on at least one axis, though two is better. This of course requires some power, some mechanical and electrical systems which cost more and are subject to breakdown, and some controlling logic. Another method is to use pseudo-mirrors (I say pseudo because they only have to reflect the light energy, not the image, so there are different priorities in design of the mirrors) made of polished metal. These have the advantage of concentrating more light to the cells, but this advantage can be wasted if you reach the light saturation point of the photovoltaics, or heat them up too much. These mirrors require mechanical and electrical systems, but are also subject to interference or even damage from wind.

    Then there's the charging system: photovoltaics rarely put out the voltage you want, and that voltage varies with the temperature and the amount of light hitting them. Pretty much every electricity consuming device has its own expectations in terms of voltage and current, so there needs to be some power electronics involved, along with controlling logic. And as soon as you get to some decent current, these become large and expensive pieces of silicon.

    Don't get me wrong, I am a big fan of photovoltaics, I am just trying to show that, if you want to do it the right way, there's a lot more involved than putting a PV cell on your roof and connecting it to the power leads of your embedded system. I learned all this when I was building rooftop server for a wifi mesh networking system using MIT's roofnet software. The problem is, not much funding was available.

  20. Gumstix SBC by tigeba · · Score: 1

    If you are in the market for something like this, try a Gumstix (www.gumstix.com). They are 200/400 mhz xscale boards with 4 meg of flash and 64 meg of ram, running 2.6 linux kernel. They have 2 serial ports, usb client, bluetooth, MMC card, GPIO. I believe the new versions with ethernet and CF slots are getting close to competion. They are also TINY and consume around 100mw going full blast (at least that is my experience). They also provide a toolchain and everything needed to build your own kernel. Very cool.

    Check them out!

    1. Re:Gumstix SBC by setagllib · · Score: 2, Insightful

      You'll have a hard time convincing anyone in-the-know to run Linux on anything that isn't x86. Its strategy towards porting is "convince the rest of the kernel it's an i386 and work like that", which fails on a lot of systems which are fundamentally different from i386, even if they have some things in common (ISA isn't the same everywhere, for instance).

      NetBSD is portable in the Right Way. It actually abstracts architectures, busses, etc. completely, with nothing i386-specific leaving the i386 world (to combat confusion: ports that are LIKE i386 but wouldn't involve the same kernel and installation are exempt - e.g. Xen which doesn't need any hardware drivers but is still based on i386).

      The same clean code and clean design that allow this kind of abstraction lead to a generally clean and stable system. NetBSD's worst stability problems occur only in device quirks which haven't yet been fully understood (you'll notice Linux has the same quirks but the hacks around them are usually done earlier, since Linux contributors don't care if something is a hack or not). Where the hardware is non-quirky, the system simply does not encounter problems. Simple as that. You could say the same for Linux to some extent, but for Linux, achieving stability is all about quirks for everything, even where it's not needed [see the first mention of i386 'emulation' above]. That's why the Linux kernel is an order of magnitude larger than NetBSD's but does not have the functionality to justify it.

      I ended up typing a lot in response to very little, but the point is: Linux is marketable, so it gets around, NetBSD is proper, so it gets installed over Linux by anyone in their right mind.

      --
      Sam ty sig.
    2. Re:Gumstix SBC by tigeba · · Score: 1

      I think you entirely missed the point of my post, which was that the Gumstix is a fantastic, featureful and inexpensive SBC. I'm sure the (in your opinion) superior NetBSD would run on it just fine, considering how common the arch is.

    3. Re:Gumstix SBC by setagllib · · Score: 1

      Yeah, there's that, but you also advertised its 2.6-ness and toolchain to build a new kernel. So it's not like the software on it didn't interest you :)

      --
      Sam ty sig.
    4. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      You'll have a hard time convincing anyone in-the-know to run Linux on anything that isn't x86.

      That is complete crap. I'll give you a couple of people who use Linux on non ia32. SGI, IBM, plenty of ARM products, Alphas used to be (not sure if they are still sold).

      So I assume you'll say you know better than all these people because you have some ancient SGI MIPS machine that nobody cares about and Linux doesn't work on it, right?

      How about you show a few examples of people in-the-know who run NetBSD. Yourself doesn't count.

      Its strategy towards porting is "convince the rest of the kernel it's an i386 and work like that", which fails on a lot of systems which are fundamentally different from i386, even if they have some things in common (ISA isn't the same everywhere, for instance).

      Sorry, Linux is ported to more CPU architectures than NetBSD, including architectures without MMUs and PPC64 (which doesn't have pagetables), neither of which NetBSD can handle.

      So give me a single example of something that is i386 specific (or even i386 centric for that matter) in generic kernel code, and I'll eat my hat. If not, you are just a clueless troll.

      The same clean code and clean design that allow this kind of abstraction lead to a generally clean and stable system. NetBSD's worst stability problems occur only in device quirks which haven't yet been fully understood (you'll notice Linux has the same quirks but the hacks around them are usually done earlier, since Linux contributors don't care if something is a hack or not).

      I don't think you are in a position to say that Linux kernel developers don't care if something is a hack or not. You're nowhere near in their league. But give me some examples of these so called "hacks" (clue: a workaround isn't a hack by definition).

      Where the hardware is non-quirky, the system simply does not encounter problems. Simple as that. You could say the same for Linux to some extent, but for Linux, achieving stability is all about quirks for everything, even where it's not needed [see the first mention of i386 'emulation' above].

      I'm sorry, but you are simply wrong about this.

      That's why the Linux kernel is an order of magnitude larger than NetBSD's but does not have the functionality to justify it.

      The Linux kernel has an order of magnitude more device drivers, supports more CPU architectures, has orders of magnitude more filesystems (including proper journalling filesystems), can be configured to run in 2MB of RAM, or scale up to 512 CPUs and terrabytes of RAM, thousands of disks, hundreds of PCI busses...

    5. Re:Gumstix SBC by setagllib · · Score: 1

      http://bsd.slashdot.org/comments.pl?sid=135621&cid =11322387

      Enough said.

      --
      Sam ty sig.
    6. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      No, that isn't "Enough said".

      Regardless, that post doesn't resolve any of your points. Linux's ISA drivers may have i386isms, probably because many were written before it was ported to such a wide array of CPU architectures. That said, Linux's moden driver layers are far more effective and portable than anything FreeBSD has got (probably comparable or better than NetBSD).

      But regardless of this, it isn't core code, which is what I asked you to provide examples of to back up your claim:

      Its strategy towards porting is "convince the rest of the kernel it's an i386 and work like that"

      So again I ask:

      Show a few examples of people in-the-know who run NetBSD on !i386.

      Give me a single example of something that is i386 specific (or even i386 centric for that matter).

      Give me some examples of these so called "hacks" in drivers (clue: a workaround isn't simply a hack by definition).

      And tell me what functionality would justify Linux's size if more CPU architectures, order of magnitude more drivers and hardware support, orders of magnitude more filesystems, including proper journalling filesystems, scalability from 2MB RAM to 512CPUs and 8TB RAM, tens of thousands of disks, hundreds of PCI busses, proper hot pluggable devices and CPUs, and better performance doesn't justify its increased size? And who are you to say what justifies increased size anyway?

    7. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Give me a single example of something that is i386 specific (or even i386 centric for that matter).

      This is supposed to be qualified with: "in the core kernel code"; show me an i386ism in the core kernel code.

    8. Re:Gumstix SBC by setagllib · · Score: 1

      It being written aeons ago does not justify it staying outdated. NetBSD replaces old code. If Linux can't do that (however far from the 'core' as you seem to love, which is completely irrelevant because hardware support - the issue being discussed - is never at the core) it's always going to have these little things holding it back. You can stick on all the wonderful clustering and file systems and whatever else you want, but it won't support hardware to its fullest, at least not without dirty hacks or redundant drivers.

      If you even watched the changelogs you'd see all of the mangling needed to get some drivers working properly. I'm not even going to bother looking through it all if you won't; I've told you of something and you told me it's not true. Unless you do a full audit of Linux and prove that it does NOT use any hacks, you can only assume it does. This is logical. In software, assume the worst until you prove otherwise.

      The one thing that sets apart NetBSD from other systems is the necessity for code quality, not just as a 'nice thing to have'. If it was really that shoddy, it wouldn't be here. It also wouldn't be praised by anyone who's worked with its code as the cleanest, because you just don't make that shit up.

      Try to develop a port for both systems, like the developer whose comment I posted, and see for yourself.

      --
      Sam ty sig.
    9. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      It being written aeons ago does not justify it staying outdated.

      Yes it does, because if nobody has the hardware or motivation to test it, it doesn't always get updated. If somebody cares, it will. The fact that nobody has updated it means that nobody cares. Hence why would you anyone bother?

      OK, that one embedded guy may care, but hey, if he's not going to contribute his code back, why should anybody help him?

      And he may be quite right that there are old drivers out there in the tree that are i386 only. Linux did used to be i386 only itself. It is not anymore, and current driver interfaces are good enough to support more architectures than NetBSD currently supports from the same source tree.

      NetBSD replaces old code. If Linux can't do that (however far from the 'core' as you seem to love, which is completely irrelevant because hardware support - the issue being discussed - is never at the core) it's always going to have these little things holding it back.

      The issue being discussed in *this* thread was that you said Linux treated everything like an i386, which it doesn't.

      If you even watched the changelogs you'd see all of the mangling needed to get some drivers working properly. I'm not even going to bother looking through it all if you won't; I've told you of something and you told me it's not true. Unless you do a full audit of Linux and prove that it does NOT use any hacks, you can only assume it does. This is logical. In software, assume the worst until you prove otherwise.

      Yeah right, nice cop out because you lost the argument. This is how it goes: the person to make a claim must provide proof. No two ways about it.

      The one thing that sets apart NetBSD from other systems is the necessity for code quality, not just as a 'nice thing to have'. If it was really that shoddy, it wouldn't be here. It also wouldn't be praised by anyone who's worked with its code as the cleanest, because you just don't make that shit up.

      It seems you make a lot of shit up because you still have provided ZERO proof for ANY of the numerous wild claims you have made (in this and other threads).

    10. Re:Gumstix SBC by setagllib · · Score: 1

      Found your i386 centric Linux core:
      http://www.wasabisystems.com/pdfs/Linux_or_BSD.pdf
      The Linux memory management system is designed around the three-level MMU available on Intel x86 processors. For these and similar processors, this works extremely well. However, systems with other MMU designs are forced to suffer the complexity and performance impact of making the underlying hardware appear to function like a three level MMU system. In many cases this requires code to perform specific low level hardware access (for example to flush TLBs) to be scattered throughout the kernel. NetBSD, by comparison, has a cleanly designed pmap abstraction that provides a well-defined interface for the high level routines to perform virtual memory related operations. Each processor's low-level pmap code can then implement the data instructions and algorithms best suited to its MMU.

      We're done here.

      --
      Sam ty sig.
    11. Re:Gumstix SBC by setagllib · · Score: 1

      Since you couldn't find it yourself (or were too scared to), here: http://www.wasabisystems.com/pdfs/Linux_or_BSD.pdf

      I'm now using this as a reference to Linux unportability because it DOES support what I'm saying, unlike your dumb ass.

      An example you'll find elsewhere in this sub-thread:
      The Linux memory management system is designed around the three-level MMU available on Intel x86 processors. For these and similar processors, this works extremely well. However, systems with other MMU designs are forced to suffer the complexity and performance impact of making the underlying hardware appear to function like a three level MMU system. In many cases this requires code to perform specific low level hardware access (for example to flush TLBs) to be scattered throughout the kernel. NetBSD, by comparison, has a cleanly designed pmap abstraction that provides a well-defined interface for the high level routines to perform virtual memory related operations. Each processor's low-level pmap code can then implement the data instructions and algorithms best suited to its MMU.

      Will you please just die now? Give up: Linux is not abstract among architectures. It's a hack job and it's a damn miracle it even boots. It can't sustain many ports in-tree, that's why there are mips-sources, sparc-sources, alpha-sources, and so on. Fixes and improvements are not properly shared or cross integrated, so you never know how F'd up your Linux is compared to others. NetBSD has every port in one tree and it IS abstracted, so improvements affect everything related. If you had even once run it you'd notice. I have run Linux and NetBSD and I know the difference. It's nice to have corporations who research operating system architecture back it up.

      --
      Sam ty sig.
    12. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Dude, are you serious? You really think page tables are an i386 centric feature? To start with i386 has 2 level page tables (PAE uses 3 though).

      Let's see what architectures have page tables... alpha, arm, cris, frv, i386, ia64, mips, parisc, s390, sh, sparc, sparc64, x86-64....

      One architecture that doesn't use page tables is the PPC64. There is NO requirement there that the kernel makes the hardware appear like it has 3 level page tables... in the case of the PPC64 port, there *are* page tables which are backed by a hash based MMU.

      The page table _abstraction_ is used for the basic backing store of virtual page status. This is a damn *great* abstraction because it is what 90%+ of hardware has been using since the 70s.

      Quoting IBM, the POWER5 is 95% as efficient on AIX as it is on Linux. That is pretty impressive considering AIX is basically tuned exclusively for POWER. Let's see NetBSD match that. Oh wait, it doesn't run on POWER.

      You are an imbicile. You are just parroting what you hear from random sources on the internet with no clue yourself. You are a troll sir. Go and pass the second grade before you open your mouth again.

      i386 centric feature my arse. Retard.

    13. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Since you couldn't find it yourself (or were too scared to), here: http://www.wasabisystems.com/pdfs/Linux_or_BSD.pdf

      Apply for clues here. They'll be half price for the next hour.

      Hey, those guys are either lying or don't know that much about Linux, because they also say:

      With Linux, however, device driver code must be reworked for every new architecture.

      Bzzt.

      As a result, NetBSD supports fifty one supported architectures from the same source tree.

      Funny these guys count their ports based on system architecture, which isn't a very impressive measure of portability. CPU ISA is the better indicator. They fail to mention Linux supports nearly 50% more CPU ISAs than NetBSD from their source tree.

      Under Linux, many devices require explicit hardware information before they can be used.

      Bzzt.

      Moreover, there is no unified source tree for all architectures

      Bzzt.

      I could go on, but I can't stand reading this drivel.

      Will you please just die now? Give up: Linux is not abstract among architectures. It's a hack job and it's a damn miracle it even boots.

      You are a troll, sir.

      It can't sustain many ports in-tree, that's why there are mips-sources, sparc-sources, alpha-sources, and so on.

      Funny it supports more architectures than netbsd from the main source tree.

      Fixes and improvements are not properly shared or cross integrated, so you never know how F'd up your Linux is compared to others.

      Err what pipe are you smoking on? I suppose you think you're an insider to the devlopment process now? Oh no I forgot, you just make claims without proving them. You have still not proven a single one of your retarded claims you retard.

      NetBSD has every port in one tree and it IS abstracted, so improvements affect everything related. If you had even once run it you'd notice. I have run Linux and NetBSD and I know the difference. It's nice to have corporations who research operating system architecture back it up.

      Baaahahaha! Are you joking? You think Linux doesn't? IBM, Intel, Sun, Orcacle, NEC, AMD, Sony, Fujitsu, Dell, HP to name a few you may have heard of.

    14. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Wait, I just have to laugh at you a bit more.

      BAAAA HAAAA HAAAA! You stupid driveling moron. Are you fucked in the head or just have absolutely no idea about the most basic CS principles that have been around for decades? If so, why would you try to make statements up about Linux and pass them off as fact? Surely you know someone who actually knows this (pretty basic) stuff will catch you out.

      BAAA HAAA HAAA HAA! You retard, you actually thought page tables were an i386 centric feature? That's the funniest thing I've heard all day. Thank you for the good laugh.

      PS. Don't bother posting anything anymore. I'm tired of arguing with you as you obviously have a sub-chimp IQ. Oh and if I ever see you posting anything again where you try to pass yourself off as being an expert, I'll just link back to this.

      Amazing. Truely amazing. You really have no idea and yet you still insist on arguing. You are a sad, sad excuse for a man. If you are under the age of 10, I'll excuse you.

    15. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Quoting IBM, the POWER5 is 95% as efficient on AIX as it is on Linux. That is pretty impressive considering AIX is basically tuned exclusively for POWER. Let's see NetBSD match that. Oh wait, it doesn't run on POWER.

      Uhh, should read: POWER5 is 95% as efficient on Linux as it is on AIX.

    16. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      The page table _abstraction_ is used for the basic backing store of virtual page status. This is a damn *great* abstraction because it is what 90%+ of hardware has been using since the 70s.

      That wording is slightly misleading.

      The page table solution for the virtual -> physical memory management problem is a great structure for a number of reasons I won't go into (look them up in a basic CS101 book).

      It is used by 90%+ of hardware *because* it is such a good solution to the problem. And it is used by Linux for the same reason.

      It doesn't mean the hardware *has* to use page tables to be able to run efficiently under Linux (see IBM's POWER). The fact that most architectures *do* have page tables is a bonus.

      If you think they are an architecture specific feature, you are a moron. But we have already established that earlier, so I won't elaborate further.

    17. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Dude, are you serious? You really think page tables are an i386 centric feature? To start with i386 has 2 level page tables (PAE uses 3 though).

      But I must just remind you that the 3 level page table system was put into place years ago, before the 386 had PAE. It was for the Alpha, actually.

      And now it is 4 levels anyway, for the x86-64 (although other architectures can also use for levels, including ia64, ppc64, s390).

    18. Re:Gumstix SBC by setagllib · · Score: 1

      You've made your point a dozen times, get over it. Nobody cares what you know about architectures and page tables. I quoted a source that could have been better chosen. I did NOT say page tables are an i386 specific feature, I assumed that the pdf therein implied that implementations were different and NetBSD was abstract where Linux compensated for differences with layers. Excuse me for trusting third-party resources. I can see you're the authority because you personally reviewed the low level hardware documentation for all of these processors and then the source of NetBSD and Linux relevant to them, and so are completely pure of 'reading disease'.

      Get over it. I'm not anti-Linux, I'm anti-(people who say that Linux portability is a work of art and NetBSD is irrelevant). Hell I have four machines powered up in this room and three of them are Linux. That doesn't mean it's a work of art; it happens to be convenient in ways NetBSD won't be until Gentoo Portage or something like it is available. We all have our reasons.

      --
      Sam ty sig.
    19. Re:Gumstix SBC by Anonymous Coward · · Score: 0

      Nobody cares what you know about architectures and page tables.

      Oh, I know you don't care, because you don't care in the slightest for facts. You just say whatever pops into you head at the time. Let me quote you:

      "NetBSD is portable in the Right Way. It actually abstracts architectures, busses, etc. completely, with nothing i386-specific leaving the i386 world (to combat confusion: ports that are LIKE i386 but wouldn't involve the same kernel and installation are exempt - e.g. Xen which doesn't need any hardware drivers but is still based on i386)."

      I did NOT say page tables are an i386 specific feature

      Seems you are a liar too. Let me quote you again:

      "Found your i386 centric Linux core:
      http://www.wasabisystems.com/pdfs/Linux_or_ BSD.pdf
      The Linux memory management system is designed around the three-level MMU available on Intel x86 processors [...]

      We're done here."

      Excuse me for trusting third-party resources.

      Oh yeah, like you did with your "new SCSI subsystem in 2.6" post. You troll. You don't go around saying how bad Linux's portability is and how many hacks it has in it because you're trusting some "third party". If you do, that means you're a troll.

      and then the source of NetBSD

      No, I have not. The difference is, I don't go around making outrageous and false claims about NetBSD for fun. I'll be quite happy to put trolls like you to bed though though. Every day of the week.

      You stupid hypocritical imbicile. Listen to what you are saying then have a look at your post that started off this whole thing. Then listen to your "I'm not anti-Linux, I'm anti-(people who say that Linux portability is a work of art and NetBSD is irrelevant)." bullshit. Fuck off.

  21. Condensation by MarcQuadra · · Score: 1

    Alright, what you need is a plastic case, like a millitary ammo case. You drill a few small holes for the cables (don't expose the ports!) and some silicone putty to seal it up. Before you deploy this, make sure it's air-tight by removing the board and heating it in the oven for a few minutes and seeing if the expansion of gas is contained (it will 'pop' when you open it; if it doesn't, try again. Put the board back in, toss a few big packs of silica gel in there (you can buy milspec silica packets that are the size of an iPod) an seal it up.

    Don't use moving parts, and keep the thing from vibration and direct sunlight when it's out in the field.

    Oh yeah, use a silicone lube on the rubber gasket that lines the opening, it will last a LONG time that way.

    --
    "Sometimes, I think Trent just needs a cup of hot chocolate and a blankie." -Tori Amos on Nine Inch Nails