Slashdot Mirror
AMD To Stop Production Of 486, 586 & K6 Chips
Dear VersaLogic Customer:
This letter is being sent to alert you to an important change in the long-term availability of several VersaLogic products. Please read it carefully.
AMD, the supplier of CPU chips that are used on many of our products, has notified us that they plan to re-tool the production line that currently produces 486, 586 and K6 CPU chips. AMD needs to use their Fab 25 facility to produce a different line of products and will stop production on these CPU chips on June 28, 2002. The CPU chips produced by this facility are used in our VSBC-2, VSBC-6, VSBC-7, Panther, VL-686-2, and VL-586-1 products.
This decision by AMD, with whom we have worked closely for many years, is a major blow to the embedded computer market. It is very surprising that their long-standing dedication to the embedded market has taken such an abrupt turn. As recently as October 2000 they announced new processors (the K6-2E+ and K6-IIIE+) and assured us of their continuing long-term support for the embedded market.
Please note that this decision by AMD does not mean that they will immediately halt production or that these CPU chips will be in short supply. Normal production of these chips is scheduled to continue through June 2002. Last-time-buy orders can be placed in June for delivery of the chips in late 2002 and early 2003.
VersaLogic management has been hearing rumors of this possible change (various versions of it) over the last few months and has been working closely with AMD to avoid this radical change in their direction. We prepared for the possibility that their decision would ultimately be to issue an end-of-life notice. Now that the decision has been made, our focus will be on assisting our customers with planning and migration issues over the next 12-24 months.
Although this change is not immediate, each customer must look at the long term impact that this announcement will have on their product usage. In some cases this will mean placing an end-of-life purchase order with VersaLogic to continue delivery of the current product even after the AMD chips have been discontinued. For others it may involve qualifying new products, or using Intel Tillamook versions of our current products, for the current application. Tillamook versions of most impacted products will be available before year end. For further information please see the roadmap and migration information on our web site at http://www.versalogic.com/support/rdmp/rdmp.asp or contact us directly at info@versalogic.com.
Again, this change is not immediate, but planning steps should be taken now to assure a smooth transition in the future. We stand ready to support you as needed to make this transition as easy and painless as possible. "
I've gotten a couple of these e-mails - full test of the e-mail is pasted below.
This is a test of the AMD emergency broadcast message. This is only a test. If this had been an actual email, you door would have been kicked in by federal agents, your AMD CPU's would have been confiscated, and you would have been arrested for violation of the DMCA. We now return you to your regularly scheduled email.
As long as they don't stop production on their flagship processor....
Probably because they're fast enough to do the task required for many embedded applications, and they require significantly less power than a Pentium class chip.
"The invisible and the non-existent look very much alike." -- Delos B. McKown
A lot of people have said things like "x86 isn't used that much for embedded programming anyway," and that's clearly false.
Generally, the trend in the embedded -- specifically automation and control -- markets, is to move from expensive and non-forward-compatible ASICs and SH processors and the MIPS series to x86 processors. Why, you ask? Because x86-based PLCs can be programmed using a standard compiler, instead of a special cross-compiler like the Green Hill Compiler (which costs a lot).
Did I mention the cross-compilers for SH/MIPS/etc cost a *lot*?
By using x86 one COTS compilers. Conceivably, if you're using COTS equipment for the buses (standard UART, etc.) you could compile applications and OS using VC++, gcc, Turbo C, etc. x86 for embedded/PLC might seem braindead, but the cost savings outweigh the programmer's headache. This is especially true if you're running in real mode and don't have to worry about segmented memory (no matter; most embedded x86 programmers just initialize the segment registers to the same value and the offset registers to MAXINT and in doing so, get a flat memory model)
In addition, x86 is the primary target of VxWorks, UC/OS, and other off-the-shelf operating systems. The advantage to using 3rd party operating systems is, you don't have to spend time and money designing your own to find it incompatible along your product line -- especially if the low end of your product line is an SH processor and the high end is a pentium III. By using x86 for the embedded market, you can cash in on standardized, third party OSes and not have to worry about backwards/forwards compatibility.
So now that I've finished ranting about how x86 is a big cost saver, let's talk about why 486-K6 is important (from AMD's point of view). Let's face it. You couldn't use the athlon to power ANY industrial or consumer appliance -- unless you're talking about an oven. My athlon 1400 hits 55C and that's WITH a FOP38 cooler and four case fans. Air flow issues I may have aside, this is clearly unacceptable for thermostat controls, or assembly line mechanisms, or automotive controls, or space shuttle computers, or smart refrigerators, et al. By having a low-power K6-II (my laptop uses a K6-II/400 and it runs pretty damn cool) one can get optimum performance at a low cost, using very little power. Combine the "low cost/low power draw/reasonable performance" benefit with the "standardized OS/save costs on cross-compilers" benefit and you can see why x86 is compelling for embedded control applications.
Personally, if this is true (I've seen no announcement from AMD proper, only from this forwarded memo), I think it's going to be a big hit for AMD and other companies alike. It's going to be a big hit for AMD because they're going to lose money on a big, if unsexy market (embedded is FAR more important than PCs now, and in the future will be more so). It's going to be a big hit for embedded programmers because Intel will have a monopoly on the x86 embedded market. As more and more managers decide to move from SH/MIPS/Zilog/whatever to x86 so that they can cross-compile from COTS compilers, they're going to be pushing more money into Intel's hands. Intel can then reasonably do some serious price gouging, claiming "it takes extra effort to keep these 386E, etc plants open" even though the plants are a 'sunk cost' in terms of capital.
Well, the men in white coats are ranting... and they have blue faces?
(three tones)
Three Step Plan:
1. Take over the world.
2. Get a lot of cookies.
3. Eat the cookies.
I remember the 34010. A fun little chip. I particularly liked the host interface registers!
TI also provided a library for those who did want to use the graphics capability. It came in source code form. This library had errors in it (it wouldn't work right if compiled with the optimizer on).
They also had a font library and a CCITT Group IV Fax library.
I remember the bit addressing. It took a while to get used to opcodes being on mod 8 addresses. The other thing that the 340x0 (I used the 34020, too) had was that you could specify two specific word sizes (anywhere from 1 to 32 bits), so you could move, say, 17 bits in a single pop. Weird...
Fascism starts when the efficiency of the government becomes more important than the rights of the people.
Lucent's Wavepoint II wireless access point uses AMD 486's (586's in the newest models i think) for it's processor. Along with the Apple Airport and the similar Lucent Residential Gateway. I'm not certain but I believe the Linksys cable modem routers use them too. This announcement is alot bigger than many might think right off the bat....
No, but many are still in production, and available at your local music store or audiophile shop.
The oldest tube still is mass production is the 6SN7, which is an octal-base dual triode first introduced in 1939. It's often used in old TV sets as a horizontal oscillator, but it was originally designed as an audio tube - a more primitive version of the venerable 12AX7.
In contrast, try to find me a new production 6BK6. Good luck. (Though I have several dozen new old stock 6BK6s, they've been in their original boxes since the 1950s.)
Old parts are easier to buy than many newer ones. Last year's IC is hard to find, but something that has been in production for 10 years often has enough usage to keep it going.Uhh... Unijunction transistors, like the 2N1671, were in *very* popular production 40 years ago. They were discontinued about 20 years ago. At this point, while they're not easy to find, I know of at least 4 places that will sell me a NOS 2N1671 UJT in mass-production quantities. But they're expensive.
By contrast, even Radio Shack sells 2N2222s, which originated about the same time as the 2N1671.
Specialized parts have short lives, general purpose parts stay around.I think Intel saturated the IC market well enough with their processors. I'm sure that new 486s will be available for some time to come. Maybe you'll have to rework the board or design an adapter that will let you put a BGA chip into a ZIF socket; who knows. But I'm sure you'll be able to do it one way or another.
Fire and Meat. Yummy.
Only when absolutely necessary. The R&D cost of re-engineering a product dwarfs the cost of saving a few pennies by moving to more modern, high volume hardware.
So this decision will cost some companies in R&D. That sucks!
No... no, wait, I'm an R&D engineer... yah, AMD! ;)
If you were blocking sigs, you wouldn't have to read this.
If you're doing something like controlling a VCR, a sprinkler system, TV channel changer, a thermostat, a stereo, or numerous other tasks a processor like that might be a bit too powerful.
There are more powerful things than that powered by a 386-class processor.
I used to work at a company that made high-end infrared equipment - the type of stuff you see on Cops or see featured on those shows on Fox (you know, "World's Wildest Police Chases 9"). Their high-end gimbal unit, the balls that are mounted on helicopters, used 386 processors.
And these units were _certainly_ magnitudes more complex than a VCR remote or stereo.
You've forgotten the fundamental purpose of a company: to make money. These aren't charitable organizations. They live by supply and demand.
The old standards get in the way of the new, and for that reason alone should they be phased out.
The old standards are far cheaper than the new (one of the first posts is a joke about 486s going for $3, but he's making a good point indirectly). Without the profit margin of using the older technology, the money for research and testing of new products would barely exist for many companies. They must strike a balance between offering new, expensive products, or older, cheaper *and sufficient* technologies. Not every embedded product needs the speed of an Athlon.
Developers: We can use your help.
Did I mention the cross-compilers for SH/MIPS/etc cost a *lot*?
Hiya. I work for MontaVista Software. We publish a linux-based cross-development kit which targets (among other things) several varieties each of mips, SH, ARM, PPC and x86. We use the GNU tools -- gcc, gdb and kin -- on all of these, and in most cases few bugfixes and changes are needed.
In any event, the tools we work with and publish are free. Getting a full copy of our cross-development kit (which comes with lots of nifty target apps, and good phone support, and whatnot) is liable to cost a fair bit, but (being that it's mostly GPLed) you get the rights to make your own modifications, redistribute and so forth.
Interestingly enough, not one of the targets we've got here that I've tested on has been based on AMD chips. We had Athlon support internally some time back (some of the "embedded" systems we work with are actually fairly beefy) but it's been dropped for whatever reason. Draw your own conclusions....
In any event, quality cross-compilers for odd platforms aren't really all that hard to come by.
Curmudgeon Gamer: Not happy
Companies who have spent millions of dollars designing and testing an embedded device running on a rock solid, low power 25Mhz SX 486 don't want to go through the whole process again to upgrade their boards (different pins, more power) or even to put in a faster 486 (you might have to actually cripple your application to keep its execution speed constant).
The cost of the CPU isn't really the issue (you often pay more for older, slower chips!), it's the associated re-development cost that keeps the demand for old chips going.
If you were blocking sigs, you wouldn't have to read this.
In the desktop PC market, AMD had basically replaced the K6 with Duron, and they'd done the same with the recent mobile Athlon and Duron products for laptops. I suspect they just weren't seeing a lot of demand for the older classes of processor anymore (or at least not enough to justify a fab anymore), and decided to let Intel service the low-end embedded market instead. A year-plus for a transition period isn't too bad, though companies making medical products that use embedded AMD would want more (I believe changes like that have to be certified, and that takes time/money).
Given that AMD has only a fraction of Intel's resources, that's probably a smart move on their part. Spend your money where the opportunity for a return is best. Interestingly, the embedded market can make money (at least a little higher-up) - that's pretty much what's kept PowerPC cranking along all these years. It's popular in cars, printers, and networking equipment, to a much greater degree than Apple buys them. I think Intel still makes i960s, too - for that purpose.
-- Josh Turiel
"2. Do not eat iPod Shuffle."
I recall the Zilog Z80 and Motorola 68000 had long lives after being licensed over to Hitachi.
Maybe Hitachi could start making K6s.
Treatment, not tyranny. End the drug war and free our American POWs.
See my user info for links.
What, me worry?
NOW how will I realize my dream of constructing a Beowulf cluster out of 486 SX/25s? If I get a few hundred more it'll be JUST AS FAST as my Athlon 500!
- A.P.
"Remember when the U.S. had a drug problem, and then we declared a War On Drugs, and now you can't buy drugs anymore?"
New technology is great because it enables you to do things that were not feasible before, but using a K6 where a 486 will do is just a waste of resources.
I think people here who are saying "Big deal, it's how the industry works.. Old CPU's stop getting made" are missing the point.
The "embedded" world is a bunch of companies producing devices that are usually small, lower powered, small production runs, and generally get made for alot longer than most electronics you're used to.
The last company I worked for (Midway Games) made arcade(coin-operated) video games. For a brief time, I worked with the group called "Wavenet". Wavenet was an idea to link arcade games up in arcades all through the world, to allow real-time tournaments. The first game that was tried was Mortal Kombat 3. MK3 used a really weird processor called a 34010 from Texas Instruments. (Weird in that it had *BIT* addressable memory, funky graphics opcodes built in that we never used, etc) However, the game designers pretty much pushed the CPU to its max before we had a chance to make it a networked game. There wasn't enough RAM, CPU, or ROM (for networking code) left to do it, as well as this board didn't have an ethernet output on it to connect it up to the router.
Midway ended up designing a tiiiiny little board (running a small embedded OS that just translated game commands into TCP/IP and vice versa) that plugged into an expansion connector on the MK3 board. It had an Ethernet controller, some ram, more ROMs for the networking code and a 386SX CPU made by AMD on it. Why not use a Pentium, or Pentium Pro? (which was the newest CPU out at the time)
Cost. Right now, you can get 386SX CPU's for a couple of dollars.
Power. Compare the latest generation of 386 CPUs to even a slow PII. HUUUGE difference here.
Board space. The embedded 386's are a little bigger than an american nickle. Pentium class CPU's... well... are big.
Longevity. When we bought these, we got committments from our suppliers that the CPU would be around for at least X months/years. This is REALLLLY important to us. If we're going to spend a ton of cash designing a board based around a CPU, we don't want it to disappear next month when something better comes along.
Had the embedded world not existed, and we had to use a faster/newer CPU, the board cost would have doubled, it would have been a bigger board(again more $$), We likely would have needed to put a bigger power supply(or played tricks with regulators), and then had to redsign the board every time the trendy chip got unpopular. All for horsepower we didn't even need!
Take a look here. Intel is still supporting and selling 80186 CPU's, for embedded controller uses.
Many many companies depend on slower CPU's for things. I don't know if it's still true, but at one point nearly every computer-controlled traffic light system sold used an 80186 CPU. Intel(?) came up with a "hardened" version of it that tolerated extreme cold and extreme heat. Companies that produce products like that are even happy paying double price for an old CPU that can do that, than installing air conditioners and heaters in every traffic light box.
The embedded CPU industry is a place where normal PC economics do not apply. It's not unheard of to pay extra for a part just because you know it'll be around for 10 years, instead of a cheaper(sometimes better) part that will go away as soon as it's not trendy.
While I don't know the specifics of this deal, it sounds like AMD is breaking their previously announced EOL(End Of Life) dates. This is quite likely going to piss a lot of people off who built their product around one of these CPUs.
Probably because they're fast enough to do the task required for many embedded applications, and they require significantly less power than a Pentium class chip.
Absolutely. I own a Garmin II+ GPS receiver, which is powered by an Intel 386EX processor - an unglamourous CPU, but one that does the job very nicely. Similarly my Palm III is powered by 16MHz Dragonball processor (68000 derivative) - nothing special these days, but serious grunt 10-15 years ago.
I may be showing my age here, but I first started computing at around 10 years of age on Z80 boxes (TRS-80/System-80: where 48k RAM and lowercase displays were a luxury!) I am still amazed at how good some of the software was on such limited hardware. Conversely, I see how crap some of the current software is and think "WTF"?!)
Competitions like the Perl obfuscated code contest are cool, but I'd like to see some contents based upon the best program than can be squeezed into an "antique" box like the TRS-80 Model I - that would really separate the hackers from the script kiddies...