Reading the LSI Strucutred ASIC fiasco thread has made me think. People are saying the FPGA revenues are going to grow, so....
Which markets are FPGA heading into?
I mean, at the moment there's Comms, Medical, Military, Consumer.
Where are they going next?
Automotive I guess is coming, as is aerospace. You could put the two together, as control electronics.
How about seeing them in a PC?
What are your views on the matter?
Aerospace posses issues with ionizing radiation.
When there's an advantage of reconfiguration ability over static asic massproduced at low price. A possible app could be builtin reciever for television, modem etc.. that can be adapted fast to new codecs post manufacture.
I want a board with a few of the largest Spartan3e chips, a little SRAM, and a PCIExpress 8x slot and controller for under $300. Until that happens, FPGAs are not in the consumer market in my opinion. All the current consumer boards lack one necessary feature for general purpose acceleration and coprocessing: datastream bandwidth. Tinkering with gates (like one would do with current Digilent starter kits) should not be considered a consumer product; that is in the hobbiest arena.
I've worked with the FPGAs in the military some. They use it mostly for image de/compression and de/encryption -- the obvious uses for them. Bandwidth is a huge concern on every FPGA application I've ever worked with, especially with the military. I just wish that thinking would propagate down to the consumer level. Somehow the graphics card companies seem to have a grip on it, while other general coprocessing hardware seems to have been skipped over lightly.
John C,
Automotive is happening (big) right now.
Latest (can't say which) luxury car has 10+ Xilinx FPGAs in it.
One car.
10+ FPGAs.Replaced all the microcontrollers.
Why? Because the microcontrollers can not be maintained for ten years, whereas the car maker can always buy any future version of our FPGA, and put their old VHDL/verilog into it. Maintaining stock for all cars sold for replacement assemblies for ten years was driving the maker broke (pun intended).
Consumer is going gangbusters. LCD TV's, plasma panel TV's. Why?
Because ASICs are a bad investment. The actual number of TVs sold for each country is small compared to the overall numbers. Each country is just slightly different. The TVs change every six months. So making TVs for Taiwan, China, Japan, Korea, Vietnam, Cambodia, Thailand, etc. etc. is a huge headache. And if the ASIC doesn't work, the risk is enormous.
PC's?
Already there, but it is small compared to everything else. We are in video display cards (I have one in my home PC, which can drive two displays at once, and the mouse will go from one to the next...). We are in high end accelerator cards, instrumentation cards, etc. Still like to see more business here, and it will happen with PCI express which is just starting to ramp now.
Military/Aerospace
Huge growth here. Why? Because ASICs are really pretty much dead for this market. Not because of cost (which helps), but because of reprogramability. The "mission" changes, and thus if the entire jet fighter is "soft" or "firm" it can be reprogrammed...
And there is the software defined radio retrofit of all military forces for all allies and their partners (read the whole world, pretty much) and the homeland security need to retrofit all public safety (police fire medical) radios to interoperate with the military...
Embedded systems
Everywhere there is a processor + stuff. A 20 billion $ market, which we are able to contribute to. More than 40% of it is Power PC based. So we play well here, and add a lot of value with our EMACs, APU interface, 405PPC, and thousands of programmable CLBs.
'Extreme' DSP
Anytime you can't solve the problem with a DSP processor. Or when you need peripherals for a DSP processor (you then use a FPGA instead). Another 20 billion $ market which we play in the top end of. Cell phone base stations, medical scanners, so on and so forth.
Looks like there is growth (for us), and lots of it, in our future.
Of course, that is our perspective, which is from where we sit, and what we see happening. Others will have their view.
Austin
pd....
-snip-
Both there now. Account for them as you will. We have a Aerospace/Defense/Automotive Division today.
Which is why we offer the QPro series, and have onging research into solving all the problems in these applications.
-snip again-
Already in them. The sets are programmed just as they are leaving (for country, etc.).
Austin
Consumer items do have the advantage of volume manufacture and recovery of development costs on a few hundred board does tend to make them dearer than that market never mind thge dearer manufacture costs.
Wait a long while and maybe and our Raggedstone2 will bring what you want but don't expect it this side of Christmas. It's on the roadmap but not for a long long time yet. Broaddown3 when it releases will have some what you want and probably dearer than you want to pay.
John Adair Enterpoint Ltd. - Home of Raggedstone1. The Low Cost Spartan-3 Development Board.
I'm surprised you didn't mention that quite a few FPGAs, including Xilinx FPGAs, have been roaming around the surface of Mars for the past two years. And while a few mechanical parts are starting to wear out, the FPGAs are still going strong.
See
Altho the company seems to have ceased trading, there are a few of us who have units. It uses Spartan xc2s200's, one to implement a northbridge and one as a memory controller and graphic engine. A xc95144 is used to boot them and provide memory control. NOTE this is not a windows machine, but is a RISCOS
-- webmaster@tankstage.co.uk Iyonix PC
Duane,
Yes! Exciting, but not very high volume. 12 XCV1000's. Times 2.
Aust> Aust>
?!? - I'm hoping this is just Austin's enthusiasm, because if the Luxury Car vendor really believed this, that is truly scary.
Would YOU drive off, in your luxury car, in 10 years time, where they had 'just respun' the VHDL code, thru new tools, into the newest FPGA ( _and_ assuming that newest FPGA _AND_ PCB will physically fit at all !? ).
This is nothing but a lawyer's feeding trough....
-jg
Jim,
Hee hee.
Just the facts, Jim. When we see the PO for hundreds of thousands of chips, we have to believe they are doing something with them. Believing what they told us is one option.
How is this use any different from the latest version of Windowz controling the police car's computer dispatch system?
(also true)
There is a lot out there that is very scarey.
Do I trust FPGAs any more than I trust anything else?
Nope.
But I do trust engineers will do the best job they can given the tools....
Aust> Aust>
hmm....looking at the writing on the wall economically....
PC-mkt: no growth...expect a decline.
mil/aero: a new slaughter every year should keep this going; but overall volume is questionable....
Automotive: US mfg car-sales are off 20-40% in only a few months; Delphi in bankruptcy, Dana just filed BK, GM itself hasn't filed yet but is technically already bankrupt and is teetering on the edge of filing....Ford credit downgraded to junk...25,000 layoffs in Canadian auto-mfg....about that many just announced in US; and I think many many more to come...
The chinese are making lots of claims of millions-per-year new cars on the road; but even a touch of recession in the rest-of-world will trash the chinese economy (which is massively export-based). I.e., don't count on automotive to grow. Also, automotive generally doesn't need speed beyond uController ability; and the volumes are high; so price matters.
Medical: who can afford it any more? Medical is topping...costs are out of sight, and medical expenditure correlates with jobs (paid insurance); and jobs in industrialized nations like US, Europe, etc. are evaporating like dew in the sun...
Consumer: ha ha ha.... Real estate bubble has been pricked...sales off --45%-- in Florida over the past 60 days....San Diego similar....and virtually all economic "growth" over the past 5 years has been RE/consumer driven; NOT industrial/mfg growth.
The past few years of spending-binge were fueld by conversion of equity to cash, then to debt. No more RE apprecation, no more equity to tap....and just a 10% drop in avg. RE prices will put -30%- of ARM's underwater. Consumer-bucks are about to dry up in a serious way...
And consumer products are generally high-volume, low power, and cheap....which doesn't sound to me like a great FPGA market.
My overall impression is a flattening of FPGA sales...chip sales in general really. Again, plenty of signs of it already....
In a depression, who needs the latest PC, or another cell-phone? What you -need- is FOOD, shelter, and fuel....
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
---= East/West-Coast Server Farms - Total Privacy via Encryption =---
This doesn't make sense to me. It's my impression that, if anything, the situation for fpga's is WORSE than uC's. As far as being able to get parts for many years; and especially for having new parts work in
10 yr old sockets. Just thinking of our discussion of 5v I/O over the past couple weeks as one example of where new parts won't go in old sockets...Heck, the pinout is never even the same, when you need it to be... LOL
I will note that LCD inventory is piling up worldwide. While your technical points of the desirability of fpga's may well have merit; I'm not sure that we'll see total volumes climb.
Overall, I was fascinated to read your outlook right after posting my own views; which are driven by macroeconomic factors. It will be very interesting to see which of our scenarios has played out 2-3 years from now!
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
---= East/West-Coast Server Farms - Total Privacy via Encryption =---
ROFL....
also, a point I was going to mention in an earlier response...
Luxury car is a tiny tiny fraction of automotive sales....20k/year,
100k/year maybe.On a side-note, as someone who has restored and made profitable use of a lot of older test-equipment over the past 3 decades; it scares the shit out of me to visualize half the planet running on stored-charge chips which are -destined- to lose their code. Most especially in the most critical applications, i.e. auto, aerospace, etc.; due to the hot environment.
Whole freaking world's going to come to a grinding halt as devices start failing over the next decade or two...
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
---= East/West-Coast Server Farms - Total Privacy via Encryption =---
there is absolutely no evidence that that assumption is valid.
in fact, there is ample evidence all around us, that engineers will do whatever they're told; and that they produce -enormous- numbers of sub-optimal designs; along with many truly awful designs as well.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
---= East/West-Coast Server Farms - Total Privacy via Encryption =---
What _exactly_ did they tell you ?
It was not the target I found surprising, but the _reason_ you offered.
I _can_ think of some reasons to use FPGA on the higher priced/lower volume vehicles, but I find it very hard to believe a car companies design engineers, (or even middle managment!), truly believes they :
".. can always buy any future version of your FPGA, and put their old VHDL/verilog into it. Maintaining stock for all cars sold for replacement assemblies for ten years was driving the maker broke (pun intended)."
Then what ? Do they think that any future version will drop into the same socket ? ( err Socket ?! ) .... ROTFL.
Well, we can only hope those 10 FPGAs are going into the marketing froth interior, and not something like Airbag, Traction control, Fuel Injection....
-jg
Jim,
OK, here is the problem:
If you buy Froto's uP, they make a new one every year, and you have no choice but to switch to the new unit, or stock enough of the old units to serve your entire replacement need.
So, the car company goes to Froto, and say "Froto: attention! We want the same uP guaranteed for a long time."
and Froto says: "huh? did you hear anything? did someone say something?"..."nope, must have been the wind..."
Then the car company comes to Xilinx, and says "do you have a solution?" and we say: sure, use a soft uP (like Microblaze), and just retarget the last version of the code to the next version of our FPGA (which we maintain for a much longer time in production -- just look at the products that we have that are more than ten years old, and we still sell lots of) when you go to the next model.
Most of the IP is in c code, compiled for uBlaze, with a smattering of VHDL for interface. Not even a big change from what they do now, just more efficient because they buy fewer ICs to do the same job (special goofy peripherals are soft in the FPGA). And no strange ASICs ever required.
Car company goes "darn, we could save 10M$ a year for just one model, just in the first year!
Next trivia question: why did the new car need a xc4Vfx12?
Aust> Aust>
On a sunny day (Thu, 16 Mar 2006 17:37:03 -0800) it happened austin wrote in :
mmmm a new 8052 comes out every year. We had 'clockless' 8052 here discussed not so long ago.
Not agianst FPGA in cars, just generally scared of electronics in essential parts in cars ..connector corrosion, no brakes, engine control responding to local cell-phone tower... you read the news. Lost if discussion on that in for example sci.electronics.design.
the funniest thing I did read was the Japanese (think it was) police will get a device (transmitter) that can remotely stop a getaway car..... I was sort of thinking how long it would take before the bad gusy controlled the police car...
Now we have viruses and worms in RFID tags already,
But OK nothing beats FPGA for speed and versatility.
Metal,
Pessimists always claim they are just observing reality.
You must be working for one of those companies "circling the drain?"
Really hard to be positive about anything.
I see (our) business increasing (a lot), and I see margins that are good, and getting better, and I see new opportunities popping up all over the place. And I see the financials that support it.
You have told all of us how we will need food, shelter, and a few sticks to burn here in the not so distant future.
Austin
Xilinx has their QPRO line which is radiation tolerant for this market. Aerospace was actually one of the earlier adopters of FPGAs for DSP applications. I've built my business around this market and have completed several designs specifically intended for low earth orbit insertion. QPRO is not new, it has been around for close to a decade.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.