Slashdot Mirror
Compaq Transfers Alpha to Intel
yaba was one of many who noted that Intel is apparently buying alpha from compaq. They also plan to move to their servers to Itanium. There will be at least one more generation of the Alpha chips, but you can imagine how much that'll matter. I still like alpha chips. Behold! Consolidation!Update: 06/25 02:19 PM by H :Check out my recent story about this as well.
I wonder how this will affect AMD's use of the Alpha's EV6 front side bus. Think: AMD can't use Intel's GTL+ bus because Intel owns the patent and they can't use the EV6 bus because... erm... Intel owns the patent.
I had hoped that Compaq would aggressively market Alpha with the DEC acquisition, and would offer us a choice in the IA32-IA64 migration.
I had hoped for fast and reasonably-priced Alpha systems. These never materialized. You never even gave the architecture a chance - the marketing was nonexistent.
I've had a reasonable level of respect for Compaq equipment, but now I hear that Compaq wants to reposition itself as a services company.
Shame on you, Compaq. You are the second largest computer company in the world, but it looks to the public that you are lackeys, easily threatened and controlled by Intel and Microsoft. You could have made the market a better place, but all that you've done is make everything worse.
I guess that it's all in Sun's hands now.
At the most, the IA-64 may benefit from some of the superior parts of the Alpha (and maybe standard IA-32 CPUs) like the FPU, but for the most part, the Alpha may suffer from NIH syndrome.
At least Compaq say they will port VMS to the Itanium. So while the superior CPU may die out, at least the superior OS will soldier on.
Hang on.. ported to the Itanic? That's a fate worse than death!!
Wasn't it Sun's support that asked customers to sign an NDA about a bad hardware bug?
-Paul Komarek
Was the original name. It is decended from the BBC computer, which had a 6502, but when the designers started thinking about replacing it, they didn't like the way the 680x0 and the 80x86 handled interupts, among other things, so they decided to do their own design, I think Acorn was the name of the company, or the computer. I also seem to think they were interperted, so you could swich cpu arch with no real problems.
Dear me, the things you remember from being a Newton user.
Plato seems wrong to me today
Wonder if the Alpha 833 is going to be available at dollar stores because of this.
Result: Intel can now make ever-more pathetic CPUs on the grounds that there is NOBODY to compete with it. It has a de-facto monopoly. Everyone else is specialized into tiny niches, dead in the water, or just dead.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Until they've finished pissing about with their high end products.
When you're buying large servers you want a 5-10 year upgrade lifeline in front of you.
Good news for Sun, SGI, IBM, HP.
Deleted
Hehe, farmers would be in a better position if there were 200M more mouths to feed.
$0.02
-l
Help cure AIDS, cancer, and more. Donate your unused computer time to worldcommunitygrid.org. Join Team Slashdot!
Yes, but Intel doesn't license the Socket 8/Slot 1/(What's that new socket called?). That as I understand is why AMD went with the Alpha bus in the first place. Despite the technical advantages, the marketplace had to adapt to the new bus, and manufacturers had to create boards and chipsets to use it.
If it didn't have technical advantages, it would have been a serious problem for AMD. I assume the boards are currently more expensive because of supply and demand...
If Intel locks down the new bus, they'll have pulled the rug out from AMD. I doubt the existing agreements between AMD and Alpha/Compaq include a perpetual offer to license the bus at a reasonable cost.
Will the EV6 bus be the Socket 7 all over again? Stretched to an absurdly long life until a new technology is introduced to the market? Are there any technologies left?
Does this mean that Intel has the patent on the EV6 bus AMD is using?
Compaq bought DEC for one reason and one reason only: the service organization. DEC already had a worldwide organization of service engineers on staff, and a well-oiled machine to keep it all working. In order to tackle the enterprise market, Compaq needed that, and it was easier to buy DEC than to create it themselves.
What they really didn't want was DEC's technology. Those of you who were paying attention at the time might recall that Compaq initially told their Tru64 UNIX customers that they were going to force-migrate them all to Windows NT. This sounded good for a while, until the customers shouted back, "Screw you, we're going to Sun!" That made them back off.
Perhaps Compaq has now decided that it's time to finally let go of Alpha -- a technology that they feel is "baggage" when their bread and butter is Wintel. Itanium is clearly their desired destination. The only reason they give credit to Linux is because right now it's the only operating system that actually runs on Itanium.
It's a shame, I like Compaq's hardware -- I've always found it to be very well-built (albeit proprietary in places where it shouldn't be) -- but their dedication to the Wintel monoculture is quite unattractive.
--
Tired of FB/Google censorship? Visit UNCENSORED!
Well, VLIW/EPIC is actually very similar to RISC. The major difference is that a RISC CPU normally does branch prediction, code reordering, etc. (in real time) while VLIW/EPIC relies on a compiler to do it off-line (i.e. not it real time). Thus, theoretically, the ia64 architecture would not suffer pipeline stalls at all, but it does make compilers a whole lot more complicated. Also, relying on a compiler to do what CPU normally does in hardware, (theoretically) makes the CPU less complicated. Well, we'll see how it plays out. So far it's been a rough ride for ia64. It's what 3 years late now? And the Merced err... "Itanium" is just a testing CPU, due to be replaced by McKinly.
___
___
If you think big enough, you'll never have to do it.
the reason sun hardware is so expensive is cuz people are willing to pay that much (though, who knows why that is).
Because their support is so damn good.
Some of these have also been noted by other respondants, but let me see if I can summarize:
As for some of your other comments:
In summary, I think your conclusion that the Intel based design is the only serious contender out there is a bit overstated.
-"Zow"
Right...but for my purposes (UNIX based realtime music synthesis and DSP) the Alpha running Linux would have kicked some serious ass.
Your choice of product names gives away just how out of touch you are with current technology offerings in the way of OpenVMS and Tru64 Unix's TruCluster Software.
Tru64 Unix (as it's now known) offers VMS style clustering on UNIX. No other vendor can currently compete in this arena. Sun's Serengeti doesn't match up, and Linux isn't there yet either. The GFS (Global Filesystem) is a significant step in the right direction, but that's only one piece in the puzzle. By the time they get parity with what TruClusters offer today, TruCluster will have moved on even further. Today, TruCluster can boast a common view of the mount table across cluster members (for any filesystem type that can sit on top of CFS). In the works, for example, are similar enhancement to the process table. I don't think that kind of functionality is even on the Linux development radar yet.
Macka
you'll probably get a few more years out of your
new boxes. Compaq have sizable military contracts that depend on VMS. And they have been doing an awful lot of development for the platform lately.
K.
-
-- Proud descendant of semi-nomadic cattle-herders.
Intel does x86, Itanium and ARM - and soon Alpha.
AMD, Cyrix and Transmeta do x86 compatibility and play catchup with Intel's x86 development.
Sun (Sparc), HP (HP) and IBM (PowerPC) do their own server & workstation chips as usual - on a very low level (counting CPUs here), as usual.
According to my CPU statistics Intel is the singular heavyweight with respect to CPU development here?! So Intel aquiring one of his (very) few competitors is a Bad Thing(tm) with respect to a healthy architecture diversity.
A current theory regarding Compaq's so-called marketing efforts for the Alpha platform was that it was headed up by the same people who ran Digital's crack marketing team.
Twice?! I could only dream of hearing from Compaq marketing that many times. I've been working with Alphas since, oh, around '94-'95. I have never received a call from anyone in Compaq's marketing group that wasn't a return call to one initiated by either myself or a co-worker. Salespeople who don't know how to sell. Unfortunately, that seems to have been one of the things that Compaq received when they bought Digital. Aside from some printed materials, I have found out about new offerings in Alphaservers either from the www or from local resellers. Mostly the latter, since almost immediately after Compaq's purchase of Digital, their web sites became intensely graphical and slow to the point of being unusable, links that point to nothing, pages that don't really tell you anything, etc. You become aware that there's something wrong when you cannot view pages on the official Alphaserver, and especially the Tru64, web site using a workstation running Tru64 and the browser that ships on the installation CDs. It's so nice having to use the Intel/NT box on your desk to research Alpha/Tru64 purchases. Compaq's marketing efforts are pathetic for anything that isn't Intel-based.
Even though Tru64 has been rumored to have already been ported to the Itanium processor, the comments around the office were NOT ``Well, that's a relief!'' but, rather, ``Why wouldn't we just run Linux?'' Why, indeed!
--
CUR ALLOC 20195.....5804M
"promoted and marketed" hahaha
I own a couple of alphas.. I'm a "registered" developer... I worked on them at a previous company.
Do you know how many times I was "marketed" to? Like... maybe twice. I worked on a total of 10 alpha systems.
Compaq has done absolutly NOTHING. NADA. Heck, their marketing sites "alphapowered" was always broken. The only freaking thing they ever got right was the processor itself. You couldn't even buy an alpha online from their site.
Oh well, maybe they'll be a great discount on hardware.
Pan
I said no... but I missed and it came out yes.
Alpha has this same instruction set (Since the EV4/5 even). It does repetitive loops over numerous registers to achieve massive multiplication/division sets (useful for MPEG encoding)
It's called Alpha MVI.
Pan
I said no... but I missed and it came out yes.
You mean like Cygnus Solutions? They're a bunch loosers anyway.
Volunteers makes it sound degrading, too. I guess the United Way are all a bunch of loosers. So's the Red Cross and Catholic Charities. All Loosers
Pan
I said no... but I missed and it came out yes.
NT is at least as poor a bet as VMS. You will have no problem supporting your current Alpha-based VMS systems for many years to come, and since Microsoft is officially on a 2-4 year upgade cycle (and changing their license terms to force you into the upgrades, see many previous /. articles) it won't cost you any more than the NT treadmill.
I know of at least two VAX/VMS systems still in heavy daily use running VMS 5 (yes, they both survived Y2K without patches or problems despite the doomsayers). How many NT 3.51 systems (which is not really a fair comparison anyway since VMS 6 predates NT 3.51) are still in use? How many of them go five years without rebooting?
Start converting to linux now; take your time, and all things will converge nicely for you in five years or less. Don't bother with NT, it's just another proprietary rat-hole.
--Charlie
PS:
I admin VMS, linux, and NT, incidentally (among others) for a living.
--CTB
There've been a fair number of old VAX (P-VAX and C-VAX) as well as early Alpha systems showing up at the DuPont Surplus Asset sales lately.
For those that don't know, DuPont (Uncle Dupie to Delaware residents) is a 200 year old zaibatsu operating mostly out of Wilmington, DE, USA. The DuPont family, that started it, is both famous and infamous for their benevolent works, war profiteering, and involvement with GM and Standard Oil. The company (as opposed to the family) is famous for inefficiency and pollution.
DuPont suffers from severe beaurocratitis and consequently has been dismembering itself and selling off the pieces for the last half-dozen years. You can get old Alphas for $100 when they have them, currently there is an HSC and a VAX 8500 processor sitting around... plus some specialized equipment for torturing small animals (no, sadly I am not kidding).
Happy trashing.
--Charlie
Most commercial unix variants can use ACLs if you want them. Linux has ACL patches.
Linux evolves over time in response to the needs of its users. WNT evolves over time in response to what Microsoft tells WNT users they want...
Choice of masters is not freedom.
Alpha has a very good software x86 emulation - I'd reckon that Intel wants the rights to that - not to mention possible patents controlling the Alpha system bus that the AMD Athlon uses...
Choice of masters is not freedom.
Microsoft competes well in the desktop market, but they aren't a serious contender in the high-end server market. Do they even have a 32-bit operating system?
Intel competes well in the desktop market, but they aren't a serious contender in the high-end server market. Do they even have a 32-bit processor?
The IBM PC architecture competes well in the desktop market....
These are all statements that could have been made in the past (and probably were). They've all been superseded. I would suggest that success in the desktop market permits economies of scale that can be used to overtake the server market. On a side note, I would suggest that success in the desktop market hinges on selling to business customers. Thus Dell is eating both Compaq's and Gateway's lunches.
Next SPARC will die. It has to because the economics for it just are not there. Sun is creeping further and further behind in the bang/$ curve and they simply don't have the money required to go to the next level.
The raw facts are it costs nearly unbelievable amounts of money to roll out a CPU in today's market. (Xylinx gets bigger devices and things might change, BTW.) Compaq nor Sun have the pockets needed to roll out the basic technology that runs the industry.
Technological wizardary not withstanding, Alpha and Sparc are doomed.
-- Multics
P.S. CISC, RISC, WISK, who cares? Economics and business relationships control what is adopted and by who. I wish it were not so, but wishing doens't make it so. It is silly even to discuss it, since these two CPUs are dead from purely economic reasons.
Back before Compaq purchased DEC, Intel and DEC were in a lawsuit over Intel keeping a bunch of Alpha technical specs (including details about there technology to run 32-bit code) to help Intel (and HP) with their IA-64 work. Intel responded to DEC's lawsuit by threatening to withhold Intel processors from DEC and issuing a counter lawsuit. This put DEC up the creek without a paddle since they made the bulk of their money from Intel based systems. DEC responded by crying, "monopolist!" If I remember correctly, at one point it looked like the government was really going to get involved. It only got settled when Compaq swallowed DEC up. The lawsuits were dropped or settled out of court and the government's investigation of Intel quietly died. I don't even know if Intel even got a slap on the wrist.
So Compaq purchased DEC, sold StrongArm to Intel, sold AltaVist to CMGI, and is discontinuing Alpha and selling some of its IP to Intel. Why did they buy DEC again? It seems a lot of effort and money for DEC's services division.
-- soldack
At the most, the IA-64 may benefit from some of the superior parts of the Alpha (and maybe standard IA-32 CPUs) like the FPU, but for the most part, the Alpha may suffer from NIH syndrome.
Sad, since it's superior to just about ALL other CPUs out there.
Actually, IA-64 is neither. It is based on something called VLIW (very long instruction word) which packs multiple, simple, independant instructions into one big instruction. Thus instruction can have its parts executed in parallel, which means that the CPU doesn't have to do instruction reordering.
A deep unwavering belief is a sure sign you're missing something...
"Modern" CISC is essentially RISC. If you take a look at the architectures of ALL modern x86 CPUs, they are internally RISC, but decode x86 instructions. On Intel, these x86 instructions decode onto multiple "micro-ops" and on the P4, it is these micro-ops that are stored in L1 instruction cache. On the AMD K7, these instructions are called ROPs (RISC-like operations). On both these CPUs, the simple x86 instructions are translated directly into ROPs or micro-ops, while the more complicated ones are translated into multiple ROPs or micro-ops. On both the CPUs, the execution units only execute the micro-ops, not the x86 instructions themselves.
A deep unwavering belief is a sure sign you're missing something...
Not really. Before the P6, all x86 chips had execution units that worked directly with x86 instructions. Now, the instruction units on the P4 don't even know what an x86 instruction is, all they handle are P4 micro-ops.
A deep unwavering belief is a sure sign you're missing something...
what about the fact that ia-64 SUCKS? compaq publishes papers about how superior alpha is, then turns right around and standardizes on ia-64! i thought it was sad when compaq bought dec and when sgi dropped mips, but this is by far the worst. intel has now bought out the superior competetion and will force another of their idiotic architectures on the world. once again, marketing and momentum win over superior design and preformance.
i think it'll be interesting to see if customers attempt to sue compaq, since they've been claiming a 20-25 year life span for alpha.
ultrasparc III is a dud. it's performance is lacking, to say the least. i have never heard of any drive containing sun firmware and i don't know why one would. just cuz it's branded as a sun drive (like quantums have been branded as apple and dec drives), doesn't mean sun wrote it's firmware. the reason sun hardware is so expensive is cuz people are willing to pay that much (though, who knows why that is).
Damn. Well... there are still options but I've always been kind of partial to the Alpha. Oh well.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
What I'm more curious about is how is this going to affect AMD! Amd either licensed or bought the rights for alphas bus architecture and uses it in all of it's new processors. What if Intel suddenly decided not to renew their license, etc..
This sucks ass. I have really wanted to get a nice Alpha system, but haven't had the money yet. Now it looks like a cheap system will never materialize...
What's more, I've been working on an Alpha emulator recently, and if the architecture is going to go down the drain, then perhaps the only good implementations left will be emulations....
Say, doesn't Cray use Alphas in its T3E machines or something? I would think that that is a pretty big market...
-----
OpenVMS will be ported to Itanium (along with Tru64 and NSK).
Compaq is, according to the article, also commiting to the release of one more generation of Alpha processors. But, I think you can assume that will be the last. They will be porting their OS technology to Itanium.
I think you are basically right. As far as I can tell from the press releases, this is what's happening: (1) Compaq is cancelling development on EV8 and successors, (2) Compaq will complete EV7, (3) Intel will get all Alpha technology (CAD tools, chip designs, etc.), (4) Intel will offer Alpha engineers positions presumable on IA-64 development, (5) Compaq will completely migrate away from Alpha to Itanium.
The last one gives it away: if Compaq is not a customer of Alpha, who is? So, yes, Alpha is officially dead (with EV7). Intel's interest in Alpha is not in Alpha itself, but in the design technology.
Glad to hear you didn't jump to Intel. :-)
Read the EFF's Fair Use FAQ
Alpha engineers have been jumping ship to AMD, and Intel knows how valuable engineers are (remember, they poached Motorola engineers) to the competition.
So who is the competition? Intel already took much of Motorola's brain trust, and Motorola keeps screwing up. IBM continues to do well with PowerPC, but that's a niche market and Intel probably figures they'll take down IBM's PowerPC later.
No, the competition here is certainly AMD. With AMD's stated goal of moving into the enterprise market starting to bear fruit, Intel has got to be a bit scared. As they say, "only the paranoid survive."
Read the EFF's Fair Use FAQ
We see that in the operating system field: IBM, the old monopolist, is keeping some pressure on Microsoft, the once-upon-a-time underdog, by funding development in and lending respectability to Linux.
I'll say again what I said in my orignal post: Worrying about monopoly and the evils thereof isn't a once-and-for-all sort of thing, and we can't divide the world into the evil and the good. Intel and MS are antitrust threats because they are NOT the underdogs, and thus can smother competition with FUD and dollars. AMD and Sun can't, yet, so we don't worry about them. Yet. [Emphasis added to the "yet"s.]
See what I've been reading.
There is a reason for the double standard: AMD really _is_ different than Intel. It isn't because they're nice guys, it's because they are the underdog. Their continued success, and their very existence, is in doubt from year to year. If AMD were to buy the Alpha, that would give them some additional technology resources and another product line which is solidly positioned at the high end of the pc market where they are weakest. AMD+Alpha would be a better competitor to Intel, and we would all benefit as Intel scrambled to raise quality and production and lower costs.
Intel is already stiff, possibly insurmountable, competition to AMD. Intel+Alpha lets Intel assimilate any valuable elements of the Alpha which can overcome the NIH syndrome, and strengthens their lead in the high end, high margin pc market where AMD really needs to catch up.
Intel+Alpha = less competition in the future, AMD+Alpha = more competition in the future. This isn't because of any moral superiority of AMD, but because AMD isn't yet big enough to screw us as effectively as Intel. If AMD "wins the war" and displaces Intel, they will of course try to do the same sort of damage that Microsoft did when they "won the war" against IBM. But remember, if you're old enough, that IBM was an evil empire too, before MS cut them down to size.
Worrying about monopoly and the evils thereof isn't a once-and-for-all sort of thing, and we can't divide the world into the evil and the good. Intel and MS are antitrust threats because they are NOT the underdogs, and thus can smother competition with FUD and dollars. AMD and Sun can't, yet, so we don't worry about them. Yet.
See what I've been reading.
IMHO, it's not a theoretical limit. There are plenty of database servers today that would cheerfully use 8 gigs. Likewise for scientific and engineering programs. Late 2003 will see this market really open up, and whoever can ship >32-bit boxen will be able to collect $100s per system in pure profit. Whoever doesn't pursue 64-bit will see themselves locked out of a lucrative market. The big chip makers see this brick wall, and that's why everybody is developing 64-bit CPUs.
-- ;-)
Kuro5hin.org: where the good times never end.
He didn't say that they make the best FPU core in the world. Can I run down to Future Shop and snag an Alpha station, a copy of Windows and Office so I can read word docs? No? Doesn't sound like so good a chip to me. Sure, it can do everything faster, cheaper, and while playing a symphonic orchestra, but if I can't get my hands on it, and use it for whatever I use computers for on a daily basis, it sure as hell doesn't qualify as 'the best.'
Vintage computer games and RPG books available. Email me if you're interested.
This is spectacular. We just completed a year-long project migrating 15 years of in house applications to OpenVMS Alpha.
The only reason we actually went ahead is the BINARY COMPATIBILITY! IA-64 won't be compatible! So in a few years when our computing demands once again exceed our systems we'll have to reproduce the entire migration.
Are there any details as to who will be continuing to support the current Alpha line, retired VAX line, and OpenVMS?
I have to admit Compaq has not been the greatest benefactor to DEC's legacy, however, for the users of digital's software and hardware massive uncertainty again surrounds us.
I should have just given in when the administrators wanted NT.
The List of Grievances with Slashdot.
I wonder if this could be considered "Prior Art" against the Transmeta code morphing patents.
Dynamic recompilation has been used for a long time (see also Connectix Virtual PC, Speed Doubler, and Virtual Game Station). Transmeta's claimed innovation (I consider myself quite skilled for an IANAL at figuring out what patents say) is branch prediction using commit/rollback semantics (remember your SQL?) for the CPU registers. The prior art is something similar that was implemented in the Zilog Z80 processor (not the game boy's Sharp z80clone).
Will I retire or break 10K?
And just when you thought Alphas we're the best 64bits platform to go to if you we're anti-intel and looking for real performance....
First, Digital selling it to compaq,
then compaq who NEVER promoted alphas and canned windows 2000 pro developpement with microsoft at around build 2128 if I remember correctly, and then transfering it to intel??? Nice way to slowly die.
I remember how much the compaq representative didn't want to compare the Alpha workstations to the intel processors when he came to do a demo, he claimed it was the fastest processor, this and that, all good, but NEVER DARED to compare it to an intel for the 3d rendering performance.
They held a big bomb but never used it... they could have been one major competitor in the 64 bits arena and high-performance workstation, but they never DARED to touch intel's domination since they were selling intel boxes too. I can't beleive they've pulled a "gateway" (i.e. gateway with amiga) on the Alpha... this is so frustrating...
--- Metamoderating abusive downgraders since my 300th post.
I'd seriously think about quitting and moving to AMD. Both for ethical reasons and personnal reasons. Unless there's a closed contract like "work for us for 3 years full", I don't see why anyone would work there anymore (unless he's got the big $$ and a high level, and even then...)
--- Metamoderating abusive downgraders since my 300th post.
Go to Toms Hardware and look at the chip benchmarks.
They never benchmark the AMD with the same speed intel. For example, the AMD 1.2GHz is usually benchmarked against the 1.6GHz Intel. Just because the chip is marketted at a speed doesn't mean its equivalent to its competitors speed.
--
"That's one small step for man..."
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
Is this a desperate attempt to stay in competition with AMD? To speak bluntly, Intel hasn't released a stable chip since the P3. They keep releasing really fast chips just to keep the pace with AMD. But with the recalls of the P4, I think the chip buying community is ready to put their faith in AMD.
I think the brightest move for Intel would be to dump the P4, and just update the technologies on the alpha... The alpha isn't anywhere near the end of its cycle...
--
"That's one small step for man..."
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
If you don't have anything nice to say, say it often.
If you don't have anything nice to say, say it often.
- Ed the Sock
here's the link
http://webevents.broadcast.com/compaq/PressAnnounc ement
For those of us who are into hearing sales geeks talk.
[Note the space typo in the link is a slash problem/bug. I can see it spelled correctly in the comment box.]
Check out the Vinny the Vampire comic strip
"It is a greater offense to steal men's labor, than their clothes"
Look at the definition of RISC. First thought to look it up in the jargon lexicon but to my surprise RISC/CISC weren't included.
But then, I probably have been trolled?
Ahhh...the great dumpster continuum. Many a free computer will be found there. -- sowth (748135)
Slashdot's reputation!? Slashdot's rep is that of a bunch of college students commenting outside their experience and expertise, and it's not very far wrong. It's fun and a good read and I've learned some things around here, but it's a lot like hanging out in the Quad at Enormous State University and eavesdropping on 15 different conversations. If we had a keg we'd call it a party.
This next song is very sad. Please clap along. -- Robin Zander
from the article:
"The bottom line is: we are creating great customer value,"I hate it when they say that. It'll probably mean we can all dive into our pockets again.
notable for stealing many of DEC's technologies and getting busted.
innovator [By 1995 Palmer (Former CEO of Digital Equipment Corp) was noticing reviews of Intel's new Pentium Pro line that found it strikingly--even suspiciously--improved over its Pentium forebears. Intel itself provided the most damning hints that it had leaned on its competitors for the upgrade. "There's nothing left to copy," said chief operating officer Craig Barrett in an incendiary Wall Street Journal article in August 1996. "We're a big banana now," noted CEO Andrew Grove. "We can't rely on others to do our research and development for us."]
-- .sig are belong to us!
All your
A feeling of having made the same mistake before: Deja Foobar
That's because Intel is still trying to figure out how to make a 64-bit processor. :)
/*drunk.. fix later*/
The problem with rage is that it makes it hard to focus and say what you're really feeling.
Intel has made a fortune selling garbage. The x86 architecture is just that - garbage. CISC is dead, and has been for a long time. They couldn't make a decent SMP setup until they stole technology from a International Meta Systems (P6). The flaw in CISC is the inclusion of superinstructions - why add hardware to perform a partial arctangent when memory is cheap? Wait! I have an idea...let's add 57 new ones and call them MMX. Wait! How about 72 news ones called KNI/Streaming SIMD.
Could I do it better? No. But Alpha was truely, without a question in my mind, the finestest CPU ever engineered. It's a pleasure to work on, it's fast, it scales well, and it does out of order execution, which gives it a leg up on the Ultra Sparc.
This has caused a pretty big uproar on comp.os.vms as well. As you may or may not know, OpenVMS, which at one time was THE operating system to run (if you weren't blue) only runs on the VAX and Alpha. Well, Compaq quit making the VAX in September 2000, so this is it. While Compaq claims OpenVMS will be ported to IA-64, it's hardly comparable. There are VAXen in production that haven't been rebooted in a decade. Software in place that hasn't changed in years...and this is how it's going to end? Compaq gives away Alpha technology so they can focus on the iPaq?
These are the issues raised in comp.os.vms:
Will *every* Compaq product which is sold for Alpha VMS today ( or last week ) be ported to IA64? Will they all be available by January 2004? Do they have commitments from Oracle to meet that schedule ( Oracle appear to be "excited" about Tru64 on IA64 but didn't mention if they cared VMS would be available there )? Will all existing LP licenses be transferrable to IA64 at no cost ( to systems of comparable size)? [anyone who went through a VAX to Alpha transition will understand this question]
Will Compaq provide assistance to 3rd party vendors to move their products to IA64? Will IA64 ports be a straight "recompile and link" or will some programs require substantial changes ( eg device drivers and privileged code )?
My point was that it seems to get limited respect within Compaq. I've been told ( sorry I can't remember the source ) that it was a last minute decision to port it to Alpha and that it wasn't in the original game plan. My concern is the same thing could happen in the Alpha-IA64 transition.
FMS is another product I worry about. I understand it also wasn't going to be ported, until they realized that All-in-1 needed it. FMS was not recoded for Alpha it was just VESTed. Will it be possible to re-VEST it to run on IA64? Will it be done?
And the quote that sums the whole thing up, from Bill Gunshannon:
Of course, this means congratulations are in order for it's grandfather. The PDP-11 architecture has now not only outlasted the VAX, but also the Alpha.
foo?
"All I ever wanted was to see Larry Wall give Bill Gates a Perl necklace."
http://www.eisenschmidt.org/jweisen
From dictionary.com:
vagary n : a sudden desire; "he bought it on impulse" [syn: caprice, impulse, whim]
Intel only owned the parts of Alpha that it already owned because Intel was already using Alpha technology in its chips efore it owned any of it.
Remember when Intel and DEC settled DEC's infringement suit when Intel bought the Alpha lines? How is that a settlement? Clearly DEC wasn't terribly interested in maintaining the technological independence of the Alpha design.
That same agreement multi-sourced Alpha at Samsung, AMD, and IBM. So there was and is no danger of Intel's monopolizing Alpha.
Compaq then bought all of DEC, and ended up owning whatever was left over.
(Naturally, that sounds like an inefficiency. Compaq can't handle inefficiency. Intel is organized to mediate inefficiency and even find ways to profit from it. They build a fab for one chip partly on the premise that once that chip is done in the market they can use the fab line for less-mainstream products; they've done this for 30 years; some lines are designed knowing that their primary product--this year's desktop chip, for example--will never be enough to pay the mortgage; it's a gutsy and thoroughly pro move).
--Blair
And the problem there (IIRC) is that compiler design is very tricky, since the compiler has loads of instructions to choose from and has to get it right every time. It is supposed to be very difficult, if not impossible, to write an optimizing compiler for this architecture. It's been I while since I read about this, so they might have cracked it, but it still sounds like a real obstacle.
"What are we going to do tonight, Bill?"
www.lucernesys.comHorizon: Calendar-based personal finance
Alpha isn't going to die so easily. Keep in mind that Compaq does -not- own all of the Alpha technology. Alpha Processor Inc (www.api.com) holds quite a bit of the liscensing; and while Compaq is involved in API, so are big names like Samsung. Do you thing Samsung and other such partners are going to lose the Alpha CPU just so Compaq and Intel can play kissy-face?
- Turq - "That's TRON, he fights for the users."
Based on the above, it looks like they are providing some resources and tools, and licensing the current technology. What this means (IMHO) is that they are partnering with Intel to work towards next-generation processors (Itanium and beyond), and are helping provide Intel with additional resources to improve their 64-bit line. Compaq is, according to the article, also commiting to the release of one more generation of Alpha processors. But, I think you can assume that will be the last. They will be porting their OS technology to Itanium. The plus side to this is that you may see some of the more interesting bits of Alpha technology show up down the line in some of the Intel processors.
I wonder how all of this will impact AMD....
GreyPoopon
--
GreyPoopon
--
Why is it I can write insightful comments but can't come up with a clever signature?
so the original pholosophy of RISC was to move _all_ complexity into the compiler, and simply allow the hardware to run light like the wind.
i think first gen RISC's followed that to a 'T', but the philosophy was compromised in the super-scaler generation, because of comercial desire to retain machine code compatibility with the original generation. therefore the superscaler RISCs actually branched from the RISC track, and started providing scheduling in hardware again, for the super-scaler execution order.
i think VLIW came from a combining of the original RISC philosophy with the CISC micro-programming philosophy of providing as much as much parallelism as possible on the micro-instruction level. so: allow the compiler to gen code for the micro machine. it's a very RISC kind of idea, but with a more hardware oriented outlook.
No. RISC is a philosophy. It is Reduced Instruction Set COMPLEXITY. You reduce the complexity of a given instruction, even if it means adding more instructions over all. Hense the terms Load/Store vs MOV. The only thing RISCy about x86 is adding L2 cache and FPU to the core, and that's only due to advances in fab technology. "Modern" CISC has nothing else in common with RISC.