x86 CPU development boards 86DUINO

[Oakham]

x86 dev boards interesting ?

http://www.86duino.com/?page_id=11

http://newdemo.080.net/enclosur/European/products/7

[Jessie]

300 MHz x86 processor. I'll pass. I imagine not much would be useable on that.

While clock speed isn't a true indicator of speed barring some kind of breakthrough in IPC these things have to be slow.

Edit: no FPU. About 400 bogomips at 400 MHz (stock is 300 MHz)

[Oakham]

300 MHz x86 processor. I'll pass. I imagine not much would be useable on that.

While clock speed isn't a true indicator of speed barring some kind of breakthrough in IPC these things have to be slow.

Edit: no FPU. About 400 bogomips at 400 MHz (stock is 300 MHz)

Go back to MS-DOS 6.22, and single task web browsing

[redhawk]

It's not even a damn Intel processor that's going to be fun (just like the old Cyrix that can't float)

Richard S.

[W. H. Heydt]

Looks like it's about a Win95 or an unimpressive Win98 level system. (Granted my Win98SE box--which still runs--uses a 1GHz P-III, but that was a pretty good system when I built it.) The only thing "modern" about it is that it uses current RAM.

Edit to add... The second link claims 1GHz, so that would be a solid Win98SE level system, and could probably run Win2K and--moderately likely--WinXP, albeit slowly.

[hippy]

300 MHz x86 processor. I'll pass. I imagine not much would be useable on that.

Depends what you want to do. There are plenty of application platforms out there which use hardware far less powerful than that. As a 'Super-Arduino' I don't imagine people will be primarily looking at it being a media centre, a desktop replacement, or even desiring to run Windows on it.

I never cease to be amazed that people can't imagine anything being at all useful unless it's got a multi-gigahertz CPU and four gigabytes of RAM and can run some bloated OS that needs gigabytes of storage. I don't know how we ever made anything work in the dark ages of two decades ago.

[Jessie]

300 MHz x86 processor. I'll pass. I imagine not much would be useable on that.

Depends what you want to do. There are plenty of application platforms out there which use hardware far less powerful than that. As a 'Super-Arduino' I don't imagine people will be primarily looking at it being a media centre, a desktop replacement, or even desiring to run Windows on it.

I never cease to be amazed that people can't imagine anything being at all useful unless it's got a multi-gigahertz CPU and four gigabytes of RAM and can run some bloated OS that needs gigabytes of storage. I don't know how we ever made anything work in the dark ages of two decades ago.

Standards and expectations have gone up. As has OS overhead since then.

It is more expensive than the R Pi with lower performance. I guess if you had an old x86 app and didn't have the source these would be useful. I don't think I need x86 to blink some LEDs with windows as the OS.

[mahjongg]

Yuck! Vortex SoC's the least capable peripherals integrated x86 implementation you can imagine (except perhaps an intel 80186).
Based on SiS CPU's (Silicon Integrated Systems, remember them?) disappeared from the market before 2000.
Even then they could not compete with other x86 CPU's.

Used in very old PC/104 based embedded systems, overpriced and underperforming. Only of any use if you cannot avoid using an x86 CPU (ie. if you -must- run Windows )

[Jim Manley]

Why, when I was a shiny new engineer, we didn't have no fancy Megahootz and Jiggaflops, we cranked our machines by hand, and we LIKED it that way ... we LOVED it! Then came along all those troublemakers with their hoity-toity valves and toobs that had to be replaced every eight hours, on average (well, it's their own danged fault, they were using ~15,000 of the hot, not-so-little buggers, and with a MTBF of about a year ... do the math for how long before any device failed).

Next, the discrete transistexuals (or whatever they called 'em) crowd thought they were the cat's meow what with the "tiny" (well, compared to a valve/toob), lower-power, and much longer-lived thingies. They were all terribly prejudiced though, involving all sorts of bias, and the next thing ya know, the Japanese started building 'em so small they could fit in the palm of yer hand (what? oh, yeah, them cute little radios, too).

Suddenly, biff, bam, boom, the bias was gone and all of the circuits wound up integrated - I'm surprised the resistor color code is still in existence after all that brouhaha! There were the crazy bipolar ones, and I don't know who let those Schottky folks in the door, but eventually, the MISFETs or MOSFITs, or whatever they were called became all the rage, making plumbing out of computing with everything including the silly kitchen sink, drain, gate, etc. How the heck do ya control a current with a gate, anyway? It would leak somethin' fierce, I tell ya!

There was a lot of static about mountain goats, or rams, or somethin', and then the next thing ya know, everything had to be dynamic! What the heck does a thing ya spin with the tyres on yer vehicle have to do with logic? Anyways, I lost track of what was goin' on once they started putting a whole solar system on a chip and they said even dip was too much for that, and that makes no sense whatsoever - how do ya get dip onto a chip the size of the Oort cloud? At least I think it was the Oort cloud, they kept talkin' about it like it was the Second Coming fer gosh sakes. And what's with Apples, and Raspberries, and Pies suddenly being your father's moustache? That makes about as much sense as giving a beagle a black bone!

I tell ya, it's all just crazy talk, that's what it is! Now, why did I come in here? Has anyone seen my glasses? Last time I knew it, I had put 'em up on top of my head ...

[Lob0426]

The 80186 (80c86) is still used in many products as an imbedded controller. NASA and other designers of satellites, landers and explorers used it as their primary controllers. Of course this is a very advanced 80186. Like 40Mhz. Also each one costed about $250,000 a piece as they are hardened against radiation. They are also only taken from the center of the die.
http://www.cpushack.com/space-craft-cpu.html

In the last decade or so they have been transitioning to 386+387 processors in these spacecraft. The Hubble telescope was upgraded to a 486.

Next month they could upgrade the hubble to an iPhone 4 for $49 plus a two year contract, this would octuple the computing power. But there is no USB support in the Hubble .

But 300Mhz for general computing is pretty pathetic.

[Oakham]

Seems like the Vortex is a bit of an odd ball: http://en.wikipedia.org/wiki/Vortex86

[W. H. Heydt]

Why, when I was a shiny new engineer, we didn't have no fancy Megahootz and Jiggaflops, we cranked our machines by hand, and we LIKED it that way ... we LOVED it!

...says the man who *does* hand-crank a computer.

[LemmeFatale]

Standards and expectations have gone up. As has OS overhead since then.

And speed seems to have gone right down, leaving some modern OSes only as productive, or in many cases less so, than their decades-old counterparts. Or perhaps that's just my experience?

I think I'm with Jim on this. Or something.

[W. H. Heydt]

[quote="LemmeFatale"
I think I'm with Jim on this. Or something.[/quote]

You should see the computer he was alluding to... It really is hand cranked.

[LemmeFatale]

I know.

In all seriousness, though, now I'm done with my quipping about modern OSes, I can see the use in these x86-based boards. They're just not any uses that I have myself. I'll be interested to see what applications people put them to, though.

[hippy]

I can see the use in these x86-based boards. They're just not any uses that I have myself. I'll be interested to see what applications people put them to, though.

The main appeal is undoubtedly to those who have an X86 background and are looking for something familiar which doesn't mean having to learn non-X86 or require major porting efforts for their software.

I have an automated ambient music system which is nothing more than a software MP3 player within an application which monitors the opening and closing of doors and a variety of pressure mats. I'd like to move that from a power hungry ex-desktop PC to a platform which was smaller, quieter and less power hungry but I don't want to commit to having to create everything from scratch on a new platform. A suitable PC SBC or a PC-come-Arduino platform could be ideal for this.

I also have that PC running a VB6-coded web server which serves up TV listings to my TV via a Bush IBX TV Web Browser and I'd like to keep using that with minimal effort.

None of what I do requires a super fast PC. It's running Windows 98SE on I believe a 300MHz PII with 64MB RAM, a 100MB disk with I/O interfacing via the Game Port with perfectly acceptable performance. I have stuck with the ex-desktop PC because that's the easy option; if it's not broken don't fix it. I'd like to change but it's either a matter of cost of hardware or effort involved.

If I'm developing something new I'm inclined to stick with X86 and Windows and MS programming tools because that's where my years of experience lie. I can do what I need to do and have it working in the time it would take to even work out what I needed to do on some other platform.

It's not being anti other platforms just having a lack of time resources. Something which minimises effort therefore appeals.

[Jim Manley]

I can see the use in these x86-based boards. They're just not any uses that I have myself. I'll be interested to see what applications people put them to, though.

The main appeal is undoubtedly to those who have an X86 background and are looking for something familiar which doesn't mean having to learn non-X86 or require major porting efforts for their software.

You bring up a good point that I've raised before - the major cost of a system is generally not the cost of the hardware or infrastructure software (e.g., the OS, libraries, etc.), it's the peripheral hardware and software development effort, as well as training, which can be something as informal as figuring out how to use a technology for a personal desire that's adapted from its intended use. There are other costs, though, which include the threat of virii endemic in MS software, but if your system isn't accessible from the outside (don't forget WiFi), then that may be an acceptable risk.

Then, there's the reliability issue - familiarity is nice, but it gets old the umpteenth time you need to reboot an old OS that leaks memory and otherwise becomes a dottering old fool as memory and disk space become trashed with clutter created by poor programming (in the OS, libraries, etc., if not of one's own doing in their custom code). Again, you may find that acceptable, but I wouldn't.

Then there's the issue of how you may legally use even old OS software. According to MS's infamously-worded end-user licensing agreements (EULAs), you have not and never will "own" any copy of MS software - you are solely granted a license to use it subject to MS's approval. While support for Windoze 98SE may have expired, the law school degrees (and associated loans they're still trying to pay off) of Redmond's crack team of legal beagles certainly haven't. No, they may not know how to track you down, but if you were so foolish as to make that too easy, there's ample evidence that they could make things difficult, "Capisce?"

No one can just "give away", let alone sell, old copies of such software, and copyright laws have continued to be extended to ensure that Mickey Mouse (and other, more sinister rodents) will get to live a very long and happy existence (translation: to enjoy the lifestyle to which they have become accustomed). That includes the 98SE software you're using, so you'd better start thinking about finding the receipt for it as an original, legitimate buyer, as well as the Certificate of Authenticity with its embedded shiny holographic image (it might be of Clippy, but it's really hard to tell on the bottom of a computer case that's been collecting dust on the floor for a decade-and-a-half ).

There are other issues I'm too lazy to bring up so early on a Saturday morning before heading off to run yet-another Jam, but the above should keep some posters here quite entertained for a while. Hope your pressure mats are of sufficient quality that they don't curl up on the edges like typical floor mats eventually do (assuming they're not meant to be placed under said mats already afflicted with corner-curlitis).

[Oakham]

The only reason I can see for these boards is for people who "need" to run old MS-DOS software natively rather than in emulation, due to the Windows NT family not allowing direct port access. Though saying that someone will buy one and put XP on it somehow and claim it is fantastic.

[W. H. Heydt]

I'd like to move that from a power hungry ex-desktop PC to a platform which was smaller, quieter and less power hungry but I don't want to commit to having to create everything from scratch on a new platform. A suitable PC SBC or a PC-come-Arduino platform could be ideal for this.

That strikes to heart of the whole "use high level languages for portability" issue. If code is written in a portable language that can at least be recompiled and run on a different platform/OS with minimal fuss--let alone trying for the "write once, run everywhere" ideal--then the problem tends to be far less aggravating.

Eventually...this year, next year, some time in the distant future...your Win98SE system will DIE irretrievably. What do you plan to do then? Crash project to port to a newer system? Or you could bite the bullet and start re-developing your code to run on newer, easier (and cheaper) to replace hardware with an OS that's likely to be around longer (and preferably without onerous licenses encumbering it).

This rant brought to you by an old used programmer whose first two languages were SPS (assembly language on the IBM 1620) and FORTRAN II in 1964. Finding a Monitor IID system these days is...difficult.

[LemmeFatale]

I spy a misattributed quote, there.

Finding a Monitor IID system these days is...difficult.

You've made me genuinely curious, here... Just *how* difficult is it?

[Joe Schmoe]

Though saying that someone will buy one and put XP on it somehow and claim it is fantastic.

What does this mean? Is it English?

[mahjongg]

no, the language used here is called "sarcastic".

[hippy]

You bring up a good point that I've raised before - the major cost of a system is generally not the cost of the hardware or infrastructure software (e.g., the OS, libraries, etc.), it's the peripheral hardware and software development effort, as well as training

I agree, and that's especially relevant when one has an existing system in place and it's a question of what hardware to move to. Any hardware or OS choice will have associated costs and they may be what make one option better than another.

That includes the 98SE software you're using, so you'd better start thinking about finding the receipt for it as an original, legitimate buyer, as well as the Certificate of Authenticity with its embedded shiny holographic image

The certificate of authenticity is attached to the PC case along with the installation disk.

As to 'what's bad about Microsoft'; that's been done to death probably a million times. You aren't telling me anything new. If I were worried about such things I would have moved to something else already.

Hope your pressure mats are of sufficient quality that they don't curl up on the edges like typical floor mats eventually do

They are under the carpets, bought from Maplins, have not curled up. nor have I had any problems with them whatsoever for over a decade and a half. Not sure if what they sell now (YB91Y) is exactly the same but if they are I would highly recommend them.

I'd like to move that from a power hungry ex-desktop PC to a platform which was smaller, quieter and less power hungry but I don't want to commit to having to create everything from scratch on a new platform. A suitable PC SBC or a PC-come-Arduino platform could be ideal for this.

That strikes to heart of the whole "use high level languages for portability" issue. If code is written in a portable language that can at least be recompiled and run on a different platform/OS with minimal fuss--let alone trying for the "write once, run everywhere" ideal--then the problem tends to be far less aggravating.

Absolutely and I would probably do that if I were starting from scratch today. Things were however somewhat different a decade and a half ago when I decided a Windows PC system was better than microcontrollers and hacked hardware.

It's not entirely locked down to VB6 or Windows 98 only; VB6 can often fairly easily port to VB.Net and Mono can have that running under Linux. VB6 also ports reasonably well to RealBasic ( now Xojo ) and that's cross-platform ( Windows, Linux, Mac ) though no ARM support yet AFAICT.

Eventually...this year, next year, some time in the distant future...your Win98SE system will DIE irretrievably. What do you plan to do then?

I can decide what to do when that happens. It should be easy enough for me to get the system up and working again ( no matter what needs replacing ), or I can swap to another similar system, and it shouldn't be too much trouble to get the software working on whatever Windows version is around at the time on a new X86 platform. I don't see it being a catastrophe.

Until I am forced into something I can keep weighing-up what hardware is out there, its cost, and the effort involved for various options, and decide what's best to do. So far the answer has been "do nothing".

Two years ago there wasn't a Pi so a Linux solution wasn't really any cheaper than a Windows solution. The balance on cost has swung towards a Pi but it is primarily an educational tool rather than an application platform so it may not be the best choice for the job; 'everything in a beige box, sockets on the back, I/O via PCI cards' has its advantages. Trying to replicate a 'traditional desktop PC' with a Pi isn't as easy or as cheap as one might hope it to be. Of course a PC-Arduino would likely have similar issues. I'll probably end up sticking to an X86 system as being the path of least effort even if I do rewrite the software.

[W. H. Heydt]

I spy a misattributed quote, there.

Finding a Monitor IID system these days is...difficult.

You've made me genuinely curious, here... Just *how* difficult is it?

Monitor IID was the disk version of Monitor II for the IBM 1620. The 1620 came out in the late 1950s and there were never a great many of them. Even fewer had the attached disk drive (a modified 1301, originally meant for the 1401). At the moment, I know of one "runnable" 1620 and that one belongs to the Computer History Museum in mountain View, CA. To the best of my knowledge, IBM doesn't want the museum to run it (though the museum was allowed to restore it and demo it running).

At this point, even if you could get your hands on the code of Monitor IID, it would probably be necessary to write a simulator to run on a different machine. A Pi could undoubtedly do so considerably faster, even if you used an interpreting simulator, than the original 1620. The Mod. I 1620 had a 20 usec cycle, and the shortest instruction (a NOP) took 4 cycles...making it possible to do nothing at all in 80 usec.

Difficult means in this case--for all practical purposes--impossible.

[Dietmar]

Interesting it is in any case,
because 2000 Bogomips are not bad at all,
greetings Dietmar

DMP ebox 3300
ebox 3300 from DMP is the more powerful successor of ebox 2300. Size and connectivity options are the same, just the power-on push button on the front panel is missing. Faster processor and bigger RAM gives ebox 3300 about two times the punch of ebox 2300 and therefore this is a very good platform to handle large amounts of sensor data. Most important features of ebox 3300 are:
x86 system based on Vortex86DX processor (486 compatible), 1000 MHZ, 2000 BogoMIPS, 256 MB RAM
Meteohub software and data are stored on CF card (IDE interface).
BIOS allows for restart after power fail.
Optional WLAN (WPA, b/g) and RS232 interface.
Price about $150 (without WLAN, without RS232, without CF card) at distributor from far east or US.
You can buy ebox 3300 from these distributors as listed on DMP homepage or from Norhtec (called Micro Client JrDX).