HP Introduces Final Processor in PA-RISC Family

The HP Way writes "According to an article on InformationWeek, HP announced the immediate availability of the 800 MHz, 1.0 GHz, and 1.1 GHz dual-core PA-8900 with 64MB on die L2 cache, the last member of the PA-RISC family of microprocessors. Customers with Superdome chassis can install Itanium 2 CPUs alongside PA-8900 processors."

Comments

Another one bites the dust.

[darkjedi521]

Yet another CPU architecture bites the dust in favor of the behemoth that is Intel.

Re:Another one bites the dust.

[Pierre]

Well I might be wrong here but I think that HP helped Intel design the Itanium. They've been planning this for several years I think

Re:Another one bites the dust.

[linguae]

It is truly a sad state of affairs when it comes to processors. In the PC market, now that Apple has been consolidated^Wswitched to Intel, now the x86 is the only architecture available, either from Intel or from AMD. The Alpha is dead, the PPC is now relegated to game consoles, the MIPS is relegated to embedded computers and SGI workstations, and the SPARC is also relegated to Sun workstations. There is no choice for me at all. Unless I want to shell out $5000 or more for a brand spanking new Sun/SGI workstation, scrounge on eBay to find old Alphas, or buy myself a Mac within the next year now, I will be stuck with the x86....forever.

Rob Pike said it best five years ago: there is no innovation in computer architecture and systems software at all. Everybody is focused on being "cheap" and "compatable," but nobody is focused on making an architecture that is elegant and of good quality. Nobody wants to make a new architecture that blows everything else out of the water. Nobody wants to revolutionize operating systems (I'm talking about the architecture, not the usability; Apple's doing well in the usability department). Simply put, nobody wants to try something different. And anything that wasn't Microsoft or Intel technology ends up getting destroyed. Unix was spared, but market consolidation between Unix variants and Microsoft operating systems killed many operating systems (VMS, pre-OS X Macintosh, the various Lisp operating systems, etc.). Anything new and innovative seems to be held back (for example, look at Plan 9 and Hurd).

I just wish someone would be innovative and produce architectures that advance computer science and computer engineering rather than by just "going with the flow." I want to see something fresh and new on the market. I want to have the same processor choices that people enjoyed back in the 1980s. I want to see something new coming out of those factories and those universities. I don't want architecture research to die forever. I don't want Netcraft confirming that alternative architectures are dead. I don't want Intel and AMD to be the only avenues to buy CPUs: what happens when they impose DRM on us? Intel and AMD are already in the Trusted Computing Group. Who would we run to once Microsoft demands the use of DRM'd processors in Windows 2010 and Intel and AMD begin producing their DRM-encumbered processors? We need choice, and we need change before it's too late.

Until then, where can I buy PPC, SPARC, or MIPS motherboards?

Damn

[Stevyn]

Before anyone says anything about the clock speed not being fast compared to Intel or AMID offerings, 64MB of cache is a heeeelll of a lot of cache. So all those delays from cache misses can be spent doing something meaningful...like processing.

Meta-comment

[Rebar]

Cynical remark about HP's misdirection. Outmoded sentimental longing for superior non-Intel processors. "If-only" scenario. Obvious comment about Itanium. Snarky unsubstantiated armchair prediction.

--
sig

Imagine a...

[kc32]

Athlon 64 with 64MB of cache!

You thought I was going to say beowulf cluster, didn't you?

I blame the Itanium

[SoupIsGood Food]

This is a fairly sad state of affairs... the processor family really has some legs left to it, but it was killed by HP for mostly political reasons. Itanium has never really delivered the goods, and is likely to be killed sooner rather than later by Intel, who does not know how to run a small volume/high margin performance chip line. (See: i860, i960) nor does it really see the value in such products.

Wherer this will leave HP is anyone's guess. Off-the-shelf Pentiums or Opterons can't really compete with POWER or Fujitsu's next gen SPARC designs. x86 Unix systems have largely been also-rans... Data General, Sequent(Now IBM xSeries), even Sun's new Opteron boxes are largely a side show to their SPARC business.

The Itanium, and the bone-headed wintel-centric management who pursued the pipedream of IA-64, killed off a lot of prime high-performance processor srchitectures: Alpha, Mips, and now PA-RISC. These aren't market or competitive pressures ('cuz IBM's doing just fine with bespoke silicon at the high end), but political mangement dictates that turned some premier computer science powerhouses into shambling wrecks. I mean, what the hell has SGI done in half a decade that's caused anyone to talk about them in positive terms? Nada.

This "mass extinction" of competing hardware architectures is not good for innovation. The Wintel PC is not the pinnical of hardware architectures, it's pretty bass-ackward and stone age compared to what used to be out there. Sad times.

SoupIsGood Food

Re:I blame the Itanium

[Waffle Iron]

This "mass extinction" of competing hardware architectures is not good for innovation.

The user-visible instruction set doesn't matter anymore. There's a wide variety of different architectures under the hood of the various x86-compatible implementations, and these will continue to evolve and improve. The real CPU architecture looks nothing at all like the interface presented to the programmer; this is even true for most recent RISC chips.

If non x86-compatible instruction sets provided a significant benefit, then CPUs using them would have been able to hold a substantial and lasting performance lead over the x86-compatible CPUs. But they haven't. When somebody claims that an alternative CPU architecture is beating the top-end x86 chips, it's usually just because they've slapped a massive cache next to the core. It has little if anything to do with the instruction architecture itself. The x86 instruction format is just a standardized compact bytecode that is translated to the latest features by each generation of x86-compatible microprocessor.

If you can make essentially the same progress without breaking compatibility with a huge body of software which has received so much massive investment, what good does it do to break compatibility?

Re:I blame the Itanium

[aaronl]

The point here is that the worst of all the CPU designs out there is the Intel one. Alpha, MIPS, SPARC, PA-RISC, POWER, PPC are all better designs. The reason they never really made the desktop is because they aren't Intel. This is the same rationale that lead to Windows, Word DOCs, etc being "the way".

It comes down to managers that don't know a damned thing about the tech, but making all the decisions on it. These other architectures had more growth potential, higher performance, and better overall design than any Intel chip released in the x86 line. The downside was mostly in channel cost. Since they weren't already abundant, they were expensive. If they were mass produced, they wouldn't be any different in cost than the x86 market is.

Look at how well the PPC is doing in the console industry right now. It was obviously a better choice than the x86 based chips or it wouldn't have been done. It obviously could be manufactured for the same price or less.

Two interesting tidbits. First, look into the iAPX-432 processor. Intel intended to kill off their 8-bit CPU line because in favor of that chip. It was 32bit, could do SMP, supported hot-swappable chips, and a host of other features. The 8086 was thrown in as a quick product to hold the company until the 432 was ready. Needless to say that the 432 never became popular as a result of the x86 line.

The second tidbit is that the Itanium actually needed an instruction set translator to run existing x86 apps. This layer was developed in partnership with HP. Intel *doesn't* maintain compatibility in their chips. They were trying to kill off x86 again, because it was a dead end.

Re:I blame the Itanium

[Waffle Iron]

The PA-RISC, Alpha, and Mips chip families were all way out in front of the x86 before they were put into maintenance mode by HP and SGI in anticipation of the IA-64 architecture, which never did deliver on it's price/performance promises.

That was before the x86 decoupled the inner workings from the instruction set. That's ancient history.

IBM's POWER is way out in front of the performance sweepstakes, and unlikely to be axed any time soon on the P and R series servers. Ditto the Z-series "SuperCISC" mainframe processors.

All of that is due to insane cache sizes, heavy I/O bus technology, and the hugely expensive packaging that goes with them. Cram all of that around an x86 core and you'd get similar results.

The "skin" required to make the instruction set work with bleeding edge designs like Cell just isn't worth the hassle or performance overhead.

The Cell isn't just one processor; it's a processor plus a bunch of DSPs. Nothing is stopping you from adding a bunch of DSPs to an x86 die.

The x86 instruction set is bloated and crippled

The subset of instructions that modern compilers actually issue aren't bloated, and they use up fewer bytes than space-hogging RISC opcodes. The "bloated" instructions of old are handled by a little bit of microcode if they are encountered. The x86-32 was somewhat crippled, but that was mostly worked around with tricks like renaming. The x86-64 isn't crippled.

and making the Itanium interoperable with it nearly sunk the chip entirely.

It was sunk because they thought that they could get away with shoving most of the branch prediction logic up to the compiler, which doesn't know anything about the actual runtime conditions. While the idea seemed appealing, it didn't work. An other example of how "advanced" instruction set architectures don't really buy you anything in the real world.

Allan Kay was dead right in that hardware needs to accommodate the programmer, not the other way around.

That's right, it needs to do the job as good as any alternative without making the programmers rewrite all of the software in existence.

Hopefully IBM and POWER can hold out longer.

[CyricZ]

My hope is that IBM does not make the same mistake as HP, but instead continues with their AiX/PPC combination on workstations and servers. We need variety in the UNIX market to result in innovation and improvements. With IRIX and SGI gone, Compaq and Tru64 gone, and soon possibly HP and HP-UX (there are doubts that the Itanium can fully replace the PA-RISC), the major UNIX vendors left are Sun and IBM. Frankly, that may not be enough to provide a sufficient level of innovation.

consolidation is good

[briancnorton]

Ack, the flames, THEY BURN!
But seriously, there are far too many architectures around to keep running. Fine, perhaps the elegant ones with technical superiority didn't triumph over the cruder general purpose, but I can't imagine being a developer still trying to support a dozen processors. There is market room for at least 3, and possibly 4 architectures out there, and the fewer there are, the more software choice there is for each as developers are forced to move to successful platforms.

Re:Survival of the strongest

[imsabbel]

but looking back, the alpha wasnt powered by magic either....
It was way faster than anything else, but it bought that kind of dominance by using something that now limits x86: A massive power budget.
Alphas used 80W+ back in times when 25W of a pentium2 seemed horrendious, so its not that miraculous that they got more performance out of it.

Well look at every new CPU to see if RISC matter!

[renoX]

Uh, how the PA-RISC, PPC, Sparc failures in the PC or server has anything related to the RISC concept?

If memory serves, the G5 has 1/4 the number of transistor of the P4 and it was competitive in performance.
The problem is more that even with much less transistors the economy of scales of x86 (and the intense competition between AMD and Intel), made the price very low, thus allowing x86 to compete with RISCs where it matters in the price/performance ratio, Windows and software compatibility made the rest..

Have you noticed how any new CPU is RISC?
ARM, SH, etc.. Even VLIW follow RISC conventions (fixed instruction length, load/store architecture, etc..).
So it really is a better CPU architecture than CISC but being better doesn't necessarily that you win, as shown by many examples..

Re:Did RISC really matter? Nope.

[Alioth]

What you say about IBM and H1B workers isn't true; I've worked for IBM as an H1B worker yet I do not have a Ph.D, and many of my colleagues on the project we were on were also on H1B workers. There was a critical (and genuinely rare) piece of experience we all had, but other than that we were just normal engineers.

Additionally, I was paid significantly *more* than the native IBMers because they paid me an International Service Allowance (which was generous enough I could live off it and spend hardly any of my actual salary) - so IBM was certainly not abusing the H1B system to hire cheap foreign workers because none of us were cheap.

Re:Did RISC really matter? Nope.

[Anonymous Coward]

Are people buying Pentium 4s to run DOS in Real Mode?

The frightening answer to that question is yes. There are still a plethora of programs in a variety of niche applications (machine control, point of sale, etc) that still run in real mode DOS. Many of these applications rely on hardware compatibility with the original IBM PC. That is why they still sell Pentium 4 motherboards with ISA slots.

Intel set HP up.

[team99parody]

I guess it is survival of the strongest. Intel is winning

Itanium's often laughed at for sucking; but in some ways Itanium was the most successful bluff every played in the tech industry. In much the same way that Reagan's Star Wars bankrupted the Soviet Union got almost every single competitor to fold.

Back at the begining of the project, Intel was nowhere in high-end & 64-bit computing. There was HP (PA-RISC), Sun (Sparc), Dec (Alpha), IBM (Power), MIPS (SGI). Intel wisely picked the partner with the stupidest management (Carly) to give up their competitive edge and announce to analysts that Intel's vision/roadmap is so AwSuM that RISC is dead and that they're going to follow the bidding of their master Intel for their 64-bit plan. Wall Street bought in to the story so much that almost everyone else with competitive chips folded their strong hands to Itanium's bluff - SGI spun off MIPS and MIPS decided to leave the hgh-end space. Compaq undervalued Alpha and let it die. Sun tried to become a software company and if it weren't for Fujitsu making modern sparcs, sparc would be dead.

Basically, with nothing but PR and Carly's stupidity, Intel wiped out over half of the high-end computing processor market.

Thankfully AMD had the vision to see through the bluff, and saw the opportunity for 64-bit computing that worked; and thankfully IBM didn't have someone like Carly around so they saw the value in retaining competitive advantaces; or the computing world would be pretty bleak place right now..

Re:Did RISC really matter? Nope.

[mrm677]

Sorry, you are wrong and I spoke with Patterson just a few days ago at the Internation Symposium of Computer Architecture in Madison, WI.

Patterson and Hennessy argued for RISC in the 80s before technology allowed Intel and AMD to burn 3 million transistors on a CISC->RISC translation layer. They did not forsee x86 hanging on until the mid-90s to enable this. So yes, they are wrong about the death of x86 but modern out-of-order superscalar pipelines are all based on the principles of the early RISC 5-state pipelines.

But your post claims they are failures and you are dead wrong. Among numerous other contributions, you can thank Patterson (and Randy Katz) for RAID.