Slashdot Mirror
Intel Gets PA-RISC Engineers
Doctor_D writes "Well it looks like Intel has done it again. They have absorbed more processor engineers, this time from HP. Alpha is gone, PA-RISC is going, what's going to be left? MIPS? SPARC? AMD? Crusoe? "
← Back to Stories (view on slashdot.org)
PA-RISC needed to die. It sucked. It's a damn shame Alpha's going away, but that's all water under the bridge. The important thing is that IBM is still quite alive and have produced a processor which is currently giving Intel one helluva run for their money. The Power4 chip's characteristics are nothing short of incredible. For more info check out this page at IBM Research.
Brandon D. Valentine
more engineers working on open source cores. more innovative ideas and concepts being experimented on. moores law is slowly breaking down, and as such, we need to find a way to get fresh ideas out in the open. what we need is to break away from the traditional model and start looking at some of the more esoteric ideas being thrown around. asyncronous, vliw, or a delightfully bastardized combination.
Hell, I think we should bring back the 6502 core, but that is only because I am fixated on 6502 assembly code. dont mind me, I'm a freak.
Feed the need: Digitaladdiction.net
Alpha is gone, PA-RISC is going, what's going to be left? MIPS? SPARC? AMD? Crusoe?
Zilog, of course! The greatest CPU to ever go without MUL/DIV intructions. It's a wonder it's not used in more modern computers. If it's good enough for your Gameboy and you calculator, it's good enough for your desktop, I say.
...Now If only I could find a Linux distro for my TRS-80...
Stupid like a fox!
all your brains are belong to us
Hey, can someone explain the second half of the article to me? What the hell does Heidi Roizen dressing up as a cheerleader or nurse (sketchy) have to do with the employee transfer???????
-Berj
``Dave Packard would say, `The most important thing is for a company to be profitable, period,' '' Fiorina says. ``He also was fond of saying, `This is not a democracy.' ''
It looks like Carly is turning out to be just another golden-parachuter. With an approach like that towards managing HP, she wants to turn it into a glorified Dell. Unfortunately for her, the Dell we already have is pretty darn good and commodifying HP's business is not doing what is best for the company. Don't be surprised to see her making that golden sky-dive in a year or two, while HP's share price continues to make that bright-red sky-dive we've seen for the last year or so...
When information is power, privacy is freedom.
IBMs power4?
or what about the PowerPC and its variants?
2 VERY nice architectures, and arguably the only ones other then sparc that dont seem likely to get gobbled up by chipzilla
Interesting that you refer to the Power4 and the PowerPC as two seperate architectures. They're basically variants on the same architecture; IBM makes both of them, and Motorola also makes PowerPCs. This is why you can run the same Linux distribution on an RS/6000 as an iMac.
$x='S24;r)>63/* h@<5+oZ)32"5cz';$me='phroggy'x$];
$x=~y+ -xz+\0-Tx+;print$_^chop$me for split'',$x;
The Zilog Z80 and EZ80 with embedded TCP/IP stack and web server. Kick-ass little chip. My first programming was done in Z80 assembly in the very early 80s.
I am an engineer in HP's primary PA-RISC processor development lab. I still work for HP. I still work on PA-RISC chips as do hundreds of my co-workers. Let me be clear about this: HP is NOT getting rid of its PA-RISC line or selling it to Intel or anybody else.
/. sensationalist headline and try to get the facts.
The group that was transfered to Intel worked on chipset development for some of HP's servers and workstations. Because of the economic conditions, it was deemed unfeasible to keep that group in HP, so instead of laying them all off, a deal was worked out to give them jobs at Intel. I'd say the engineers in that group are in a lot better shape than many of my other co-workers across the company who just got layed off in August.
Anyway, my point is that PA-RISC isn't dead. There are still a lot of people working on both CPUs and chipsets. We will be doing a number of iterations yet of the PA processor family before HP transitions to using IPF in the long term (we're talking years from now). Ignore the classic
Before you protest too much: They are adding Linux positions, and the money has to come from somewhere.
HP has been transitioning its processor operation to Intel for years. HP partnered with Intel to develop the IA-64 architecture. Did anyone think that PA-RISC would continue in parallel to that forever?
The world has some very serious single-source issues regarding high-end silicon in general. The fabrication lines have become so incredibly expensive as chips become more dense that most companies have given up on new CPU fab construction. And you can't make new chips with those old FABs. Perhaps we'll be lucky and there will be a revolution in microfabrication technology, but I've not heard of one on the horizon.
Another place where this hurts us is in high-end graphics, where we are down to two manufacturers.
I'd like to see more work on Open Sourced processor designs that run in field-programmable logic. This is a place where we can innovate without the expense of a fab, and then when we have good ideas that get proven, people can fab them.
Thanks
Bruce
Bruce Perens.
I remember seeing McNealy speak some years ago at a tradeshow (I don't remember exactly when or where). He explained some of the economics of CPU design and predicted that in a few years there would be only three major architectures: SPARC (Sun), PowerPC (IBM), and Intel. It's kind of creepy how his predictions have come true.
His arguement was based on financial analysis of how much it cost to develop and maintain a competitive architecture, and how much revenue each of the big players could expect from their designs. HP was doing comparitively well at the time, but McNealy figured they were using revenue from printers to subsidize their chip business, and that couldn't go on forever.
Alpha was still a viable competitor at that point, but again it comes down to volume. You might get a team of crack engineers together to design a beautiful CPU, but if you don't get the volume (or high margins like IBM mainframes) then you can't afford to keep the architecture competitive.
It is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail. - Abraham Maslow
And to further confuse matters, IBM says [IBM.com, huge graphic] that the POWER4 "includes two 64-bit PowerPC microprocessors".
I for one congratulate Intel for giving these talented engineers jobs.
Blaming GW Bush for the Iraq war is like blaming Ronald McDonald for the poor quality of food.
HP is getting out of the chip business. HP is getting out of the OS business. They are acquiring Compaq, who is also leaving the OS and chip businesses.
My question is this: What now differentiates HewPaq from say, Dell? The only thing I can think of is HP's printer business, but I don't think that alone would be enough to keep them afloat. I'm genuinely curious. What is HP doing to keep themselves relevant?
This
Well, there's IBM's Power4 architecture, Sun's Sparc3, AMD's Hammer line due RSN. I don't think the FTC has much to be concerned about with regard to IA64. In fact, Intel and HP are the ones who should be worried - Power4 passes Sparc3 and PA8700, but it blows the doors off IA64. AMD's Hammer chips will match or beat IA64 at lower cost, so Intel/HP are going to be hard pressed.
Just my opinions...
My company develops for HP-UX and solaris. We actually sell (indirectly) about 400 M $ worth of HP-UX and Solaris servers. That's a lot!
Well, there have been many in product management that were dismayed at HP's feet-dragging commitment to hteir own platform. HP has been yoyo-ing back and forth between WindowsNT and HP-UX a bit too much, and this might be the nail in the coffin, as far as our platform of choice is concerned. We'll probably standardize on Sun computers. If it happens, I must say HP really did it to itself.
Sigged!
PA-RISC is/was a great chip. The PA-8000 series was especially impressive. I'm sad to see this happen. Teaming up with Intel was probably the worst mistake HP has ever made.
Free Hans!
What exactly do you think HP uses on their servers? Doritos?
I'm a loner Dottie, a Rebel.
Sparc3 = Sparc III. WTF, Sun thinks they're the Roman Empire or something?
Intel manages to keep the prices over two times higher than an AMD with SAME performance... while I don't really care about the graphics industry right now (ATI can manage to arm Nvidia for the comming year or two, plus they have diversified fields of interrests as well to back up the company), but I am worried by Intel. In comparison, Nvidia brings you quality, performing, and rather cheap parts if you don't want to go to the top of the food chain... Intel by comparison, they are good desing (the chips don't toast, the heat sinks are easier to install than on AMD, etc) but poorly performing and WAY overpriced.
If they can manage to sell some stuff over 2 times the price of an "equivalent" AMD part, what's going to happen if AMD dies or has a major problem with the next product cycle in a year or 2 from now? Forget about crusoe, I'm talking high-end CPUs, x86 renderfarm nodes, etc...
Anyways I do my part, I buy only intel when *really* needed, i.e. when I'm instructed to, or when the programmer needs an SSE2-capable CPU to do his optimisations. Else I try to support AMD the best I can... I am about to build another renderfarm and it'll be using TigerMP an AhtlonXP processors. They need support, and Intel needs to see that it's pricing scheme is bad.
--- Metamoderating abusive downgraders since my 300th post.
Damn right! First learned assembly on the 6502, looked at the Z80 briefly, then saw the 6809... Kick ass chip. Was a work of art.
Blogging because I can...
I don't think most of them will last at Intel--even for top-level engineers, Intel runs an unpleasant and demanding working environment.
And, well, that's probably going to do a great deal of harm to HP. Because if this is how they treat their some of their best line employees, how will they treat the rest of them? HP, historically, has been an unusually good place to work. But I think this is a public sign of the end of that.
I wonder how long before all the high-tech manufacturers are unionized?
(Anonymous post because it is entirely possible that Intel keeps blacklists. It's not cowardice, it's caution, and it's going to become more common.)
Alpha is gone, PA-RISC is going, what's going to be left? MIPS? SPARC? AMD? Crusoe? "
*cough* PowerPC *cough*
IBM is still running quite hot with the power4, and motorola, though they have their heads up their asses, is assisting them with the next gen PowerPC.
yes, there's still MIPS and SPARC.. and even AMD (though they're essentially an x86 clone).. and what do they all have in common? they're not controlled by Intel, for one, but they're also all.. wait for it.. RISC
even the next gen intel chips are going risc-ish.
okay, no more martinis in the early afternoon for me...
- Entertaining Bits from the Ancient Kernel Tree
Oddly enough ZiLOG is still around (though the've done weird things to the capitalization), and they still sell Z80 and Z8 CPUs (though they prefer to call them "microcontrolers"). Much to my suprise, they even sell the Z8000, the first microprocessor to run a serious port of Unix. (These early low-end systems from Onyx and Zilog itself are long gone, of course.) And of course, Zilog is no longer owned by Exxon, which once thought the Energy Crisis was God's way of telling them to go into a new line of business.
from Compaq recently. Although I've seen nothing to indicate it is DEAD, and Compaq continues to provide support for us.
errr....umm...*whooosh* *whoosh* Is this thing on ?
Last I checked there were lots of people who owned semiconductor fabrication plants who were willing to rent out their production capacity. Fabless chip design shops are nothing new. The economics of it would be no different than any other fabless shop, only that the design would be open.
DB
Y'all forgot about PowerPC. It's not only alive and healthy, but actually growing!
Intel is pulling PA-RISC components out of IA64 in preferance to Alpha designs (SMT). Binary compatibility is endangered.
What does that mean? Like I said before, where are your facts? IA-64 has been in development for more than 5 years, completely separate from PA-RISC. What "PA-RISC components" are being exchanged for "Alpha designs?" I'm dying to know.
While there might/will be further iterations of PA-RISC, all acknowledge that the end-of-life is in sight, constantly presented on the map.
I think that's exactly what I said. Every architecture will become obsolete eventually. PA-RISC is on the way out in 5 years or more, but saying that it just got sold to Intel is incorrect. 5 years is a long time in this industry.
HP is farming out PA-RISC production to IBM, hoping that copper and SOI gives it enough boost to reduce R&D costs.
So you're saying that, because we started fabricating our chips in a better process, it means we're giving up on the architecture? Yeah, that seems pretty logical.
The porters of HP-UX to IA64 in NJ have been terminated.
...how much more evidence do you need that HP-UX/PA-RISC (and perhaps IA64) is doomed?
Don't know what that has to do with PA-RISC.
So your theory is that HP has decided to forget about the enterprise computing business entirely? I guess you're entitled to your opinion.
And what is "Sparc III"?
Oh, okay... it's UltraSparc III. But there's not been very much "Ultra" about those for quite a while, and the (late arriving, slow) III hasn't fixed this.
Well, this is the big one: http://www.theinquirer.net/14090103.htm (although this is pretty threatening to IA-64 too): http://www.theinquirer.net/13070103.htm
You seem pretty bullish about IA64. All the rest of the world can say to you is that the architecture is years late and fails to meet several design objectives (x86 performance at the very least, granted that FP is stunning). Perhaps if you had designed a better chip, you wouldn't be in this mess.
It is if HP cuts R&D efforts into further iterations - making this the last gasp of a dying line, falling into the grave a little sooner than any of us expected as the Tru64 acquisition accelerates the collapse. Oh, and you're also handing your designs to IBM, and I assume that the team that designed Power4 has now had a long opportunity to examine and be "influenced by" your work. AMD managed to implement copper in-house; Microsoft's decision to abandon OS/2 for Win95 in-house also shows a company mastering their own destiny; your choice is very telling: your destiny is controlled by others.
IA64 is the stated future of HP-UX. Why are we not to believe that Carly is terminating HP-UX with the termination of the NJ staff?
As I said, get out now. IA64 has too many problems to be a viable architecture, and the bloodbath that you are facing in competing with Hammer/Power4/Sparc will be the downfall of Intel/HP/Compaq/DEC.
It will all end in tears.
Shortening production runs and reducing retooling time and expense are a goal in *all* manufacturing plants because to create the ability to capture smaller orders enables larger profits. I don't imagine that chip fabs are different than any other production line.
In the realistic near future, I could see something on the order of a large university system (University of California perhaps?) sponsoring a chip line and having all its schools with computer engineering programs participate in the development then making a run and giving it to the EE departments to play with. That way you might cobble enough educational value out of the exercise to justify the cost and you would turn out graduates that can justifiably say that they were part of a chip design team.