
New Intel uP for Ultra-Cheap PCs 104
mircea writes "This ZDNet story talks about a new chip from Intel, dubbed Timna, targeted at ultra-low-cost PCs. It is supposed to be .18 micron technology, with integrated video. There's also mention of 'appliances based on this chip, possibly running Linux'" And the 2nd page of the article mentions AMD and Cyrix (now owned by VIA), both of whom are also preparing ultra low-cost microprocessors.
Linux toasters? (Score:1)
that means I was wrong to think Microsoft were
going to start monopolising my kitchen..
Well, I'm just glad to see cheap 0.18micron
technology
--
David Taylor
davidt-sd@xfiles.nildram.spam.co.uk
[To e-mail me: s/\.spam//]
Temna (Score:4)
I like to see what keyboards look like over at Intel, their alphabet sure seems to have a lot fewer letters than the rest of us.
Dana
Re:Linux toasters? (Score:1)
Quoting from the URL above:
What is the operating system for TiVo?
TiVo uses the Linux OS. The modified source is available under the GNU public license at www.TiVo.com/linux.
Linux (Score:1)
Now back to learning russian... DO CBNDAHR... man i need a russian keyboard for this. Throw a link to that too if you know any
Nice idea (Score:1)
nice piece of hardware for a very modest price.
I wonder if Intel consider the Celeron a success
or failure? Will they do things differenly this
time? (I suppose they think they have their clock locking technology perfected by now
One has to wory about release of specification
for things like video harware, but of course if
cost of end product is the major force behind this
then Intel cannot affor to tie themselves to
proprietary software producers.
I think this is a very interesting development
that bears watching.
Intel and Integrated Video (Score:1)
I think the `processor' will probably be very useful in handheld/portable computers... and help lower prices of laptops even more. Hey... there might be competition with the iBook
I wonder how the announcement will make the video chipset makers (like ATI, 3dfx, Riva, etc) feel.... could this mean more competition or `yet another strong-arm Intel tactic'?
did you saw this one? (Score:4)
--
http://www.beroute.tzo.com
The future is embedded (Score:3)
With embedded applications like speech recg., TCP/IP and other technologies, the embedded world is growing up to 32 bits fast. (for some reason a z80 or 8051 doesn't do TCP/IP well
Intel's move is them making sure they have IP cores (Intellectual property) that can be used in the embedded market. The high profit is in PCs, the volume is in embedded. So they'll sell 'em to PC vendors (try to starve AMD/VIA on the low end), then sell em 'forever' to the embedded world.
Chipzilla won't go down without a fight...
Integrated Components (Score:1)
That would be cool. hmmm. If only, I know something about chip fab.
Cheaper PCs magnify OS cost (Score:3)
In particular, a $75 OEM Windows 98 license that looks like "fiscal noise" when the computer cost $1500 and sold for $2000 starts looking pretty hefty when the computer sells for $400.
Of course, if Microsoft tries to eliminate "Windows 9x" in favor of the Windows 2000 pricing, that looks rather more like buying Windows NT Server at full price, this makes the OS cost more than the computer.
It's no shock that manufacturers would start looking more seriously at license-fee-free things like Linux in these sorts of circumstances...
Will It Run Linux? (Score:2)
Maybe I'm wrong, but... (Score:3)
Isn't this the path that Cyrix tried to take? Where are they now? Having worked in the past as a hardware tech for a hardware support firm, all the computers that had the Cyrix chips offering this type of integration experienced tons of problems based solely on the chip, its chipset, and its integration.
You would think that we would have learned a little bit from others that have tried it.
Also, since it's going to be so integrated, does that also mean integrated to the board? We all know what that's going to do for upgradeability, but your average Joe Blow on the street isn't. How many people are they going to have to tell, "We're sorry, but you just can't upgrade that computer... " In a market that has positioned itself MORE towards upgradeability, and more away from proprietary components, can this be a good move?
I understand that sacrifices must be made for PCs that cheap, but upgradeability should be the last, opting to lose performance over upgradeability...
BR But that's just me.
Re:Nice idea (Score:2)
>or failure?
Since AMD hasn't gone out of business yet, it obviously isn't a "success". However, they're still making it so obviously they don't consider it a "failure" either.
There's no way we would have seen a Celeron if AMD hadn't forced Intel to compete to produce such cheap but fast CPU's.
Oh YEAH! PUSH IT! (Score:3)
While some have offered PCs under $500, they have generally been with outdated, end-of-life technologies such as slower processors and CD-ROM drives.
Amazing. CD-ROM is DEAD. Incredible, considering 99.99999998% of software in the distribution market is using this obsolete media. Stop pusing so hard to make our machines obsolete! WE LIKE THEM.
But wait:
Timna will combine a Pentium processor core developed specifically for this low-cost market...
Huh? A Pentium Processor core? Isn't that a bit....DATED? We have Pentium III right now, and they're developing a chip with 5 year old technology. Hey, stick one of those end-of-life CD-ROM drives in there too.
-- Give him Head? Be a Beacon?
Target Market (Score:2)
It would seem to me that the target market for these computers are corporations that want something cheap for the average desktop worker. For these corporations, upgrading is ususally not economical in terms of labor costs, and they're going to buy a new one after the depreciation period ends.
However, I understand your worry, because these computers will end up in the hands of small business and home users that don't understand that they are trading upgradability for a lower purchase price. Think back to the old Mac LCs and Performas -- dog performance, no upgradability, unhappy users.
--
Re:Temna (Score:1)
Re:Hrmm (Score:1)
Re:Hrmm (Score:2)
Integrating everything possible on one chip usually decreases performance. AGP may be slow, but at least video cards have their own RAM, which usually gives a couple GB/s. One chip solution on the other hand is confined to the 64bit * 100MHz = 730MB/s main bus. And remember, that bus is not dedicated to graphics, so the actual bandwith is even less. Plus, Intel probably won't be putting a 15 million transistor 3D-core on this one, so you can forget playing Q3 with one of these.
Make no mistake people, this is a very low-end solution.
Re:Temna (Score:1)
"G4" has a nice ring to it too.
winTel? (Score:1)
Remember in the trial, "bill gates was livid (spelling?) when he heard that Intel were developing software for non windows platforms".. words to that affect.
Well, well, now Intel are making chips for non Windows platforms... and good thing to... especially after Intel investing in Linux etc.
So, is this a sign that the WinTel period is over? Have Intel found that they don't have to follow Microsoft like a lemming...
or Just a result of the Intel Trial, and Microsoft DOJ trail? In that Microsoft aren't going to be so pushy anymore?
or just..probably... just because such a applicanes are getting more attractive, and Intel want a part in it... (why are the tags shown - in preview with HTML Formatted option??!!)
Something is wrong (Score:2)
Are we talking cartel here, or is the demand so much greater than supply that prices bump around just under the ceiling rather than just above the floor?
Re:Nice idea (Score:1)
Happy Thanksgiving to you! (Score:1)
:-)
Why? (Score:1)
If somebody wanted a cheap CPU they could get a slightly outdated expensive CPU. If the reason for them is that it would be better than the expensive CPU's price equivalent, then why are they cheaper than the expensive CPU's performance equivalent?
Am I making any sense here?
To me it seems like it's all about marketing
Re:Oh YEAH! PUSH IT! (Score:1)
Re:The future is embedded (Score:2)
Intel has acquired one of the best embedded processor families when it got Digital's chip business: The StrongARM [intel.com], which already powers some Internet appliances-(Rebel.com [rebel.com]'s NetWinder comes to mind. So they already have a very nice line of embedded processors for all sorts of applications. Timna is obviously not targeted at the general embedded applications market. It's rather an ultra-cheap processor for inexpensive PC's and x86 compatible network appliances. I believe it will be positioned against National Semiconductor's Geode [natsemi.com], an Internet appliance on-a-chip, a great chip that's almost a complete PC. Intel has all the reasons to be scared of Geode, and VIA's own inexpensive integrated x86 processors, since Asian producers will start using them in large numbers when the Internet appliance business takes off. With a high quality, free operating system like Linux, and a cheap x86 processor combining CPU, multimedia and memory interface on a single chip, opportunities abound for companies which want to build Internet appliances.
And there are lots of processors in the embedded market which can TCP/IP quite well, along with a host of other capabilities required in today's applications. It's not like the embedded processor market is desperately waiting for Intel to come up with a line of powerful processors. The examples you give, Z80 and 8051, are not contenders in the Internet appliance market; although very advanced derivatives of these exist, which can be used for many purposes. EZ80 [zilog.com], just out from Zilog, for example, is completely backwards compatible with Z80, and has its own TCP/IP stack.
Please make sure to check Geode out. You will be impressed.
--
BluetoothCentral.com [bluetoothcentral.com]
A site for everything Bluetooth. Coming in January 2000.
USB/PCI only? (Score:3)
The Table of known working devices is growing, and the major classes of devices that aren't working yet are speakers. And it's not clear what's up with modems and NICs, which I'm seeing in stores, but no note of Linux support.
At any rate, from the recent groundswell of support, I expect that by the time the "reduced mobos" come out, there should be even better USB support than there is now. It's certainly in the interests of (SuSE|RHAT|VA-LINUX|Lineo|...) to make sure they get supported...
Re:did you saw this one? (Score:1)
How can it be less expensive? (Score:1)
Bare in mind that the chipset will have to be made specifically to accommodate these new CPU features.
This has been tried before and it has never worked and it never will. Not until the technology gets to the stage where the whole computer will be a single complex chip. Because this plan of 2 complex chips for the motherboard and one unbelievably complicated CPU is what Cirix tried ( didn't work ) and iNTEL itself failed at a long time ago ( ever herd of the 80186 ? ).
My solution ? A motherboard with sound, networking, video and even modem on a single chip ( possible with relatively little stress since it's mostly been done ) then add a "normal" CPU. simpler, faster an cheaper. Take a look at the VIA-GRA and similar motherboards. They are very close and the main transgression is windows only hardware.
somebody signed a deal to distribute Corel Linux and promptly put 150 hackers to work on Linux drivers in it's Tiwan office. guess who
Re:Hrmm (Score:1)
Re:Why? (Score:1)
For example, right now, I'm running a meager PII 350. Why don't I upgrade? Well, I've considered it, but the 350 I have now is doing everything I need it to, at exactly the speed I need. Yes, I could use faster, but I don't need faster. Everything basically works transparently, without me having to sit and wait too much while things happen. They just happen.
When things stop happening so smoothly, and/or when I start having to wait longer than what I do now, and start considering my hardware as obsolete, I'll upgrade. No, I won't upgrade to a PIII 550, or Athlon 600, but if I were to upgrade now, I'd get a PII450, for about $150.
I think the reason that most people go with these, is because, on a Wintel based machine, applications are so often ported with the "latest hardware" in mind. Software vendors try and make use of the latest and greatest processor and video card, instead of optimizing the code that they have, to make it more widely usable. Once a "user" reaches this point, in which they have to upgrade, they want something new, but don't want to have to pay for it. Also, I think they fall prey to the "megahertz war", in which they see a PII 400 and a Celeron 400, and the price difference, without knowing that a celeron 400 actually runs more like a PII 333 (just an example), and that they could pick up that PII 333 for about the same cost as a budget processor, and not have a budget processor.
Don't get me wrong, there's nothing wrong with low-end procs, I like the availability, and there are times when they are feasible for me to implement in my company. I simply prefer something designed with a slightly higher end in mind, and am not all that swayed if it's 6 months behind.
Just to reiterate, I don't know. So none of the information here has any factual basis. Just my take on the matter. If I'm wrong, lemme know.
Re:which companies care about Linux USB (Score:1)
By far the best Linux-USB supporter is SuSE, with many USB hardware vendors coming next.
Re:Something is wrong (Score:2)
A modern PC is much more complex and expensive to manufacture then a VCR or microwave.
Building modern CPU's is a hard thing to do. Some of the chips in your VCR might run at 0.1mhz. They don't need cooling. They are small and easy to handle. The rest of those chips are simple little things that cost very little and have very few abilities.
Compare this to a computer. The raw CPU speed is much higher. This forces manufactures add forced air cooling. It forces them to be very carefull about radio interference. The basic design is far more complex.
It is possible to sell really cheap computers. Sub-$100 is doable. But you'll never get all of the nifty features %99 of the computing public want, although maybe not really need. You get these cheap computers but cutting out functionality. A modem will cost you about $20 if you're talking about a WinModem. The CPU can be bought for another $20, the motherboard $15 etc. But the computer you get is a old-n-mouldy 486-class. Sure it'l surf the internet a bit but not many people want it.
Are we talking cartel here, or is the demand so much greater than supply that prices bump around just under the ceiling rather than just above the floor?
Not at all. The competition in the PC market is incredible. In the CPU market you get AMD and Intel battling away for control for starters. Intel may have more market share but they have to keep prices as low as possible to maintain the market share. Do you really think Intel and AMD would be a cartel with the way they are acting?
The same applies for every other segment of the computer hardware industry. There are many different harddrive manufactures around. Many different video manufactures. The list goes on and on.
The only thing keeping the average new computer price around $1000 is the fact that no consumers want anything less with the current technology. Most people could easily get by with less. But they are afraid they'll get a slow out-of-date computer.
At work I'd love to be able to buy older but reliable computers from reputable manufactures. But you can't. So we settle for $1000 computers when the only reason we're replacing our $200 ones is because they are becoming unreliable due to age. And the fact that before I started advising on what hardware to buy they bought a bunch of bad Cyrix machines that were made out of unreliable parts. Oh well...
Re:Temna (Score:2)
And this name is for a region here in Israel ("michrot timna") since this processor was designed here in Israel.
whois timna.com (Score:1)
Registrant:
Equinox Industries LTD (TIMNA-DOM)
25 Dunlop Ave
Winnipeg, MB R2X 2V
CANADA
Domain Name: TIMNA.COM
Administrative Contact:
Jonhston, Dennis (DJ705) equinox@MAGIC.MB.CA
(204)633-7564
Technical Contact, Zone Contact:
Bollman, Philip (PB1191) dwc@ESCAPE.CA
(204)488-3454 (FAX) (204)487-0159
Billing Contact:
Johnston, Dennis (DJ706) equinox@ESCAPE.CA
(204)633-7564
Record last updated on 19-Oct-1998.
Record created on 21-Jun-1996.
Database last updated on 25-Nov-1999 12:45:02 EST.
Hmmm, taken.... No relation to Intel, I presume...
just an Internet Appliance (Score:2)
It looks like they're using a custom web browser that is always on, so I seriously doubt they are even running an X server.
I strongly suspect that it will use proprietary hardware and software that will be useless for anything else.
AMD Has had this product for about 6 years now (Score:1)
Re:which companies care about Linux USB (Score:3)
It thus may not matter to them today, but might be of greater importance by late 2000. Note that I mentioned several companies that may not all feel they have an immediate stake in USB; they all do have an interest in increased adoption of Linux, and will ultimately be injured if availability of USB hardware gets widespread whilst Linux support remains limited.
Yes, it's evident that SuSE is supporting USB; hopefully we'll see better support come over time. I wouldn't mind seeing USB take off as an alternative to the hard-to-tie-down RS-232, and if it "takes out" ISA at the same time, I will not shed many tears.
Re:Timneh (Score:1)
``Is that a Timna processor?'' ``SQUAWK!''
Re:Linux (Score:1)
Have to keep at least one RS-232 port ... (Score:1)
RS-232 is just so widespread, keeping just one serial port would be a Good Thing (tm), IMHO.
Re:Cheaper PCs magnify OS cost (Score:1)
--
Re:Have to keep at least one RS-232 port ... (Score:2)
The heck with RS232, get a USB->Serial interface. I've seen them for parallel devices, implementing a serial interface shouldn't be much harder (May need to have a UART in the box). That way you wouldn't need a serial mux for the consoles on your Beowulf, just hook it to one USB chain. I would LOVE to see USB replace serial, parallel, simple ISA cards (modems, gameports, etc), even Ethernet. (Note: not USB audio, it is too bandwidth intensive and requires the CPU to do too much work)
I love the idea of USB networking, it seems like a great idea for home use. Even better would be the ability to see any USB device in the chain, from any host computer! I don't know if it works this way but it would be way cool! Would probably require some other tech, like Jini, to fix device contention problems.
Re:winTel? (Score:1)
Re:Celeron vs PII (Score:1)
Re:Have to keep at least one RS-232 port ... (Score:1)
Yebyen (Score:1)
Moderators, please don't mark this post down, I have no other way to communicate with AC's that ask me questions, and I know this is off topic. Thanx
Re:Can't beat the new Microfost keyboards (Score:1)
Deosyne
Re:Celeron vs PII (Score:1)
These are great value for money. I don't think a Celeron 400 runs like a PII 333 but the original poster did say it was just an example! For home users, there really isn't much to be gained in spending hundreds more on a `real' PII(I).
Re:Celeron vs PII (Score:1)
Umh..not entirely true. I have a C300A which behaves gracefully even when running 2 simultaneous instances of VMware, each with its virtual machine running in it. Under exactly the same load, i.e. 2 VMs running simultaneously, a friend's K6-2/300 becomes so slow it's almost unusable. All of the other parameters being almost exactly the same (amount of memory and swap, number of processes, etc.)
Garbage PCs you too can have one NOW! (Score:2)
I just saw in a Best Buy flyer a PC with over $600 of rebates included.
These things are a recipe for disaster. Do you own an old Pentium of 486 based Packard Bell? If you do then you know what I'm talking about. Nobody in their right mind is going to service these piecs of garbage. The upgrade market brings businesses PILES of money. If a PC is going to be cheaper to replace than to repair or upgrade why pay the money to get certified to repair them?
I repair Apples , Compaqs, and HPs. As well as upgrade them. If a new HD and 128MB of ram is going to cost more than a NEW PC why would anyone push them? These things will have abysmally low performance. So when Game X v3.5 comes out you have to buy a new $300 PC to play a $40 game, where does that leave you?
Compaq and Packard Bell have tried this in the past and if it were not for Compaq's server market share they'd have gone the same route as Packard Bell. Garbage PCs won't get the support of retailers, I'm not talking about the Suits that work for Circuit City. I'm talking about the grunts that deal directly with the customers.
I almost daily steer people AWAY from garbage, and I know that most of the salespeople that I know do the same thing. "Yes ma'am this computer is half the price of the other one but just look at how Game X runs on the two. See how jumpy it is on this one, see how smooth it is on the other one. What do you think is going to happen in 6 months when a new game you and your son want to play comes out? This computer will run it even slower than it is currently running this one."
See what I mean?
Not all of us are snake oil salesmen.
LK
Whatever happened to the WebPad? (Score:1)
Re:whois timna.com (Score:1)
i agree.. (Score:1)
> A modern PC is much more complex and expensive to manufacture then a VCR or microwave.
I agree. Those are the facts. I don't think these prices are a result of companies taking advantage of consumers by keeping prices high (even while fighting price wars?) because everyone is so mystified by their products that no one would think to challenge them.. The cause is concentrated on what it costs to produce the product.
Re:Cheaper PCs magnify OS cost (Score:1)
I'm sure that Microsoft would love to stop spending the money developing two seperate OSes that have to be compatible with each other. The problem is that a portion of their customer base (Gamers and Companies using DOS/Win3 apps) have told them that they are never upgrading to NT5. So, they figure they can go to the well one more time and milk the 'legacy' base for another upgrade. (Although, 'Millenium' might just be a downloadable service pack.)
--
Re:Have to keep at least one RS-232 port ... (Score:1)
Doesn't USB require CPU to service the bus? I can imagine that USB Ethernet would be a slooow ordeal.
Is it possible? People used to run ethernet with those Xircom parallel port adapters, so given enough pain, sure.
If the Intel things are targeted towards corporate market, they'll have onboard ethernet.
--
Re:Have to keep at least one RS-232 port ... (Score:1)
Ick, I hope not, I would think that it would take no more CPU than any other PCI device (except IDE, it was designed with ISA in mind and has no provision for non-CPU control). I tried to find more information but the closest I got was the USB Developers FAQ [usb.org], their website seems to be fact unfriendly.
Re:Have to keep at least one RS-232 port ... (Score:1)
This fits with what I've heard from others, and if a mouse takes 5% of your CPU, I hate to think what Ethernet does. (Although Intel and AMD might be happy!)
--
Re:How can it be less expensive? (Score:1)
There are lots of tradeoffs in any chip design. The idea is to not "waste" so many transistors on multiple execution units, out of order execution, etc., etc... and use the extra space for stuff that gets a big boost from the internal high-speed bus (video). The P5 architecture at a higher clock rate isn't a terrible PC for WP, etc. Alas, I'm still in P5-233 world, and I'm not suffering terribly. Further, if you can shrink the die size down, yields go up. Yield is everything in the chip making business.
Also, just because one implementation/architecture isn't successful and/or bug-free doesn't mean you can't hit a home run.
Besides, the imbedded market is lucrative too. And just for the record, the 186/188 architecture is still alive and well, with variants still rolling off the Intel fab lines. I've worked for co's that used 188 stuff because we could make a cheap product with only a hand full of parts. Just because Tandy tried to put 188's in a PC doesn't mean that the architecture is a total loss. Good for us, because of low parts count/board space, and a quick design to market solution.
Re:winTel? Sounds like they are in bed to me! (Score:1)
'It's an Intel Architecture product and it runs Windows, so it's a PC," Gelsinger said. If a PC maker or device maker were to use Timna in a device with limited function and an operating system like Linux, he said, "now you call it an appliance."'
I would venture a guess Bill Gates would line someones pocket to say that if it runs windows it a PC but if it runs Linux its just an appliance.
Its funny this machine never toasted any bread for me or did my laundry. But it fancies itself as a web browser and word processor from time to time!!
What a crock.
Re:AMD Has had this product for about 6 years now (Score:2)
Think of it this way, as you missed the whole point of the post... It isn't a new idea, it has been done before by AMD (and Motorola with their Dragonball etc) and as far as anyone here knows, the AMD K6-2+ will include an integrated graphics controller and graphics memory bus... (I doubt it, but it might)...
The embedded / one-chip-solution market is going to become huge. People will buy a PC when it comes in a black case with a small LCD screen on the front and can be stacked in the HI-FI, or embedded in the TV, etc, for less than a couple of hundred dollars. Of course, then it isn't a PC...
Of course, I believe in PCs with dedicated chips for different things. That is why the Amiga was so good in its time - dedicated graphics, sound, I/O chips etc. The all-in-one solutions are naff to me. But they will allow your mother to surf the web or whatever. I will always go for the (more expensive) dedicated stuff. Except for than damned winmodem that I bought because it didn't way winmodem on the box, so I thought it was a conventional modem... I have learnt though.
Let the processor process, the sound cards do sound (and mp3 decoding in H/W), the graphics card do graphics (and DVD decoding in H/W! Not software). You don't see software based SCSI adapters do you?
There is always a market for the crap low-end components though. I buy expensive hi-fi separates, but people wil buy $200 integrated crap-boxes with over-the-top design...
Re:USB/PCI only? (Score:1)
Yes, but practical info on Multics is getting rare (Score:2)
And the main Multics systems known to still be operational is at my brother's "office," DND: Maritime Command, [stratus.com] in Halifax, reportedly planned to be decommissioned next June.
Re:Why? (Score:1)
Undoutably one of their key advantages has been the ability to update and customize configurations dependant on customer needs.
This is also one of their key disadvantages. Try just looking at a PC sometimes. It's a virtual hornet's nest of cables and wires, most of which just serve to carry signals from one part of the computer to another. How much of the total complexity (read cost) of a PC is just to support this kind of legacy interconnects?
It seems rather obvious that most people in the world probably don't care that much about all the flexibility that a PC offers. They'd like to buy an inexpensive machine to access the net and play a few games. An internet appliance if you like. For the most part, the traditional "assemble-it-yourself" computer gets in the way of that.
Re:Linux toasters? (Score:1)
Re:Have to keep at least one RS-232 port ... (Score:1)
Intel shoud stop trying to kill it and Admit as much.
Re:Have to keep at least one RS-232 port ... (Score:1)
This is like the iMac without removable media. A great idea in theory, but everybody I know owning an iMac ended up buying a Zip. So in my opinion, Apple should have included a Zip drive (or some other RW removable media) and saved the customer the assle of buying and installing one himself.
Beside that, USB is still a pretty flaky technology. RS-232, with all it's limitation, is well understood and pretty much Plug-and-Play : most OS recognize any serial mouse without problem and as soon as you your modem is hooked up, you can send AT command.
Re:Repeat after me: The Network is the Computer (Score:1)
I admit I could live with such a setup, but I'm an exception, not the norm.
Re:How can it be less expensive? (Score:1)
How much are you willing to wager that putting the original P5 design ( Pentium before MMX, F00F bug and all
LAST POST (Score:1)
Re:Linux toasters? (Score:1)
Re:VA now employs a USB developer (Score:1)