Follow Slashdot stories on Twitter


Forgot your password?

Microsoft Bites It On 64-bit Microprocessors 373

Lots of readers pointed us to this page that says MS has stopped trying to make NT run on 64-bit Alphas and isn't even going to release any new 32-bit Alpha products. For more info on the subject, check this Netcape Netcenter article and the Ixnay WinNT on Alphastory we ran last Friday. Meanwhile, according to a short blurb in The Register, Intel has finally prototyped Merced in silicon - and it runs Linux fine but won't run Win64 at all. It looks like Microsoft simply can't deal with 64-bit architecture. Please try not to say "I told you so" too loudly to your MS-boosting coworkers, okay? ;-)
This discussion has been archived. No new comments can be posted.

Microsoft Bites It On 64-bit Microprocessors

Comments Filter:
  • What scares me is that, in light of all the problems they're having with NT, what's to keep M$ from buying Red Hat (or some other Linux retailer) and bundling crappy ports of M$ Office to Linux. I mean to them, they then get on the bandwagon, they leverage their crappy office products to take current users and installed base to Linux. THEIR Linux. They'll make Linux inexpensive, but you can bet you bottom dollar, they'll push the M$ Office bundle! And it gets them out of that nasty old monoply lawsuit!

    I don't know, maybe i just had too much coffee today? :)
  • Is it just me, or has Linux spread like kudzu in the same 5-6 year period that Microsoft has been steadily retreating? At one point, NT was supossed to run on Intel, PowerPC, MIPS and Alpha. Linux now runs on these and more, while NT is only on Intel now. When Transmeta, or whoever comes out with a new, hot chip, it seems that it is easier to get Linux on it than almost anything else, certainly easier than NT.
  • If M$ can't get it working on the IA64 arch. soon, then they'll try to make Intel change the specs, so that they can run Win2000 on it with no changes. (So there won't be any reason to buy the processor) If they can't get it working, they'll make some sort of statement on how "The general Public is not ready for 64 bit computers." simply because they aren't ready. :)
  • It's just NT 5 with a new name.

    Of course, when MS comes out with the *real* follow-on for Windows 98, their already confused naming conventions will have to undergo some big revisions.

  • That's a benefit for Microsoft and their "key partners.

    It will be interesting to see how many of their 'key partners' Microsoft will stab in the back this time. Somehow transitions like this seem to benefit Microsoft much more than any of their partners.

    Everyone must buy new apps. Everyone will want to have "64bit clean" systems.

    It depends on how long it takes for people to upgrade. If its like the last time that the PC industry went through this (the transition from 286 to 386), it was both good and bad for vendors. It is hard to say at this point if this transition will be easier or harder, and if people on the desktop side will see the upgrade as compelling. On the server side, it is hard to imagine people won't, but they may be resistant to have mismatched client and server hardware. At the time of the 286->386 migration there was a lot smaller entrenched base of existing machines.

    It will generate years and years of gluttony; several OS revisions and all-new-everything.

    That is certainly what Microsoft would like to see happen. Time will tell if it does.

    And then when it's all said-and-done nothing will have changed -- everything will look the same and performance will be back to what you'd expect.

    If everything goes as Microsoft plans, this is probably true. However, if things don't go Microsoft's way, who knows.
    As many companies as have just spent considerable sums on Y2K related upgrades, etc., it will be interesting to see how receptive they are to paying for upgrades to W2K, let alone to upgrade desktop hardware to 64 bit.

  • How expensive is a 32/64 bit mode switch? Is it feasible to compile certain parts of an application to be 64 bit, and others (less memory intensive) to be 32 bit? Is it fast enough so that I can decide bittiness on the procedure level, or should it be at the library level?

    (Ignoring for now the complications of passing pointers of different lengths between various parts of the application. There's an interesting type checking problem here- a simplified instance of existential types, I think)

  • I guess you missed the first part of the link. I'll give you a hint, it was

    No matter how reliable you consider that site, I think it disproves your theory.

    QDMerge [] 0.21!
  • is that you ken?

  • I was telling people this about MS three years ago when Merced was just a twinkle in intel's eye, then again, last year when every OS vendor in the world except IBM, Apple, and MS made HUGE noise about porting their shit to ia64. I could understand then why Apple wouldn't, and IBM, because they only had one intel OS anyway, OS/2. But I couldn't understand why MS wasn't making any noise. Oh sure, they did have whitepapers saying that they planned a 64-bit port, but it was unnatural for them to not jump on the hype bandwagon, especially with their history of FUD and Vapor - I began to see this as a symptom - that they must be seriously in trouble on the 64-bit front if they were afraid to at least issue vapor.
    Now we know.

    "The number of suckers born each minute doubles every 18 months."
  • it's not a 'mode' switch. it uses truncated pointers and kernel support for a 32bit addr. space.
  • by dillon_rinker ( 17944 ) on Thursday August 26, 1999 @06:52AM (#1724274) Homepage
    Windows 1.0
    Windows 2.0
    Windows 386
    Windows 3.0
    Windows 3.1
    Windows 3.11
    Windows for Workgroups
    Windows 95
    Windows 95A
    Windows 95B
    Windows 95 OSR 2
    Windows 95 OSR 2.1
    Windows 95 OSR 2.5
    Windows 98 (8 versions, currently)
    Windows NT 3.1
    Windows NT 3.1 Advanced Server
    Windows NT 3.51 Workstation
    Windoes NT 3.51 Server
    Windows NT 4.0 Workstation
    Windows NT 4.0 Server
    Windows NT 4.0 Server Enterprise Edition
    (Not to mention all the service packs for the above...)

    I'm sure I missed some. Does anybody have a canonical list of all the versions of Windows?
  • Well, the way things are going it looks like the PPC750 is going to be used as a LinChip on those IBM boxes, unless there are OS's that I'm forgetting which will also run on those motherboards.

    (Well, supposedly (according to Alan Cox) the WinChip is actually rather customizable and might even be able to say "LinChip". Not that I'd buy one, now that AMD K6-2s and Celery(-2)'s are so cheap.)

  • You missed the whole Windows CE line.
  • fwiw, most of my crashes (okay, lets be specific here, BSOD-en) seem to be networking and file-system related.

    Right now, I have a folder in my Recycle Bin, that when I try to empty, I get an error saying the file or directory is corrupt. I can empty all the other stuff in there, but this one folder won't go away. I know that if I format the disk, the problem will go away, but I also know that if I format the disk, I'll have to reinstall the OS (it's not my startup disk, but I have all my apps installed there, including MS Office, which has a bunch of links and crap with the startup of the OS, so if I delete that, I won't be able to boot properly)

    in other words NT sucks, and I'll be very happy when it becomes irrelevant to the business model of the company I work for, so we can get some REAL computers. It's already starting, I have a SPARC Ultra 10 PO awaiting approval. . .

    "The number of suckers born each minute doubles every 18 months."
  • I almost did when I wrote my original post, but I didn't want to get too verbose.

    My understanding is that BeOS is genuinely a completely new kernel, designed from the ground up for use with multiple processors. I don't think the Be thread/process system owes much to Unix.

    Be's window system is brand new and owes nothing to X-Windows. In terms of speed, this is a major advantage - the system definitely has a much crisper, zippier feeling than any other OS. The disadvantage, of course, is that it starts from ground zero in applications - X applications are not compatible without an X compatibility layer (which does exist; I don't know about its performance).

    However, I believe Be is posix compliant, and it definitely supports many (but not all) of the usual Gnu utilities. I would guess that porting is not too difficult for non-GUI applications. I believe GTK+ is in beta for Be, and that should make it trivial to port GTK+-based applications.


  • That shell was not supported, and IIRC had a lot of bugs, and probably would be incompatible with a lot of modern software. You're talking about something that's 4 years old. There's so much tied in with Explorer, if you're talking about IE, and Office and desktop type applications. . .

    I wouldn't recommend it. I would instead recommend fdisk, and install Linux, but, you gotta run what you gotta run at work.

    "The number of suckers born each minute doubles every 18 months."
  • Funny tho, the GUI's still SLOW.

    "The number of suckers born each minute doubles every 18 months."
  • linux allowed me to buy a $3000 laptop. That said, at least the MSCE program allows morons with no marketable skills to stay off welfare.
  • MS _will_ make Win64 work with the new CPU. But until that time, it _IS_ pretty funny ;)


  • I have it running on a 64 meg 486-66 (in my lab).

    Believe me. It's SLOOOOOW.

    My biggest lab machine right now is a dual Pentium 166 with 64 megs RAM, and I don't plan on upgrading that to W2K. Too slow. In fact, now that I think of it, I don't even have a PII. . . I probably won't be able to run W2K at all. I think I need some new equipment.

    "The number of suckers born each minute doubles every 18 months."
  • usloth can't program, but they can come up w/ CYA TLAs. beautiful.
  • Personally, I prefer NT3.51 to 4.0. It *feels* more solid and stable as well as being more stable. Probably it's that clunky Win3.1 interface that gives the 'solid' feel, compared to the Win95 interface with Start Menus flitting up and down and random popups appearing all the time.

    Anyway, I believe that between the release of Win95 and NT4, Microsoft released the new shell as an addon for NT3.51. If you ran 3.51 with the new shell, it might be a good way to get a stable Windows system with a modern interface (as well as lower system requirements than the 64MB minimum recommended for NT4SP5). I think it was freely downloadable, but I haven't found a copy on Microsoft's site (the filename was newshell.exe, I am told). Does anyone have a copy?
  • does anybody else think "Tru64" is a stupid name?

    Stupid, but marginally better than 'OSF/1'. They can't call it 'Digital Unix' since the Digital brand has all but disappeared. I think we will have to call it 'DEC Unix' - the name 'DEC' will probably annoy the Compaq marketing people even more than it annoyed the Digital marketing people.

  • 'No one will ever need more than 32 bits:' -- Bill Gates[0]


    [0] - ok, so it's fake, but doesn't this sound familiar? :)

    -fester (still wanting flash for his hpux box)
  • Oh, I messed with a beta too, and no, it didn't crash, and yes, it was very, very slow on a PII 300 with 128mb RAM.

    However, I couldn't get any of the advance features to work, because everything relied on DNS (microsoft DNS), and I couldn't get the DNS server to install. I got a lame message that "RPC Service was not available". Just a popup dialog. No event log messages, nothing else. Of course the RPC Service showed it was running just fine.
    It was an isolated network, so there was no TCP/IP addressing issue.

    However, it did run that nifty pinball game GREAT!

    "The number of suckers born each minute doubles every 18 months."
  • Well.. running NT on a Alpha. Is just plain stupid. 32-bit system ; 64-bit arch. It's like running DOS on a PIII. A wasted computed.
  • your right. there's a problem with PCI which maps into memory at 1GB.. hence limiting physical RAM to 1GB. But I think it might have changed in 2.2.11 or so to 2GB (not sure).

    Tru64 supports scatter-gather mapping. ie arbitrary PCI address spaces can be mapped on demand to arbitrary address spaces. Also, the pyxis chipset used on a lot of PCI alpha's (PC164xx) has quite sophisticated scatter-gather support, ie cut-down page tables and a tlb (similar to the way most cpu's support virtual memory). But linux doesn't support this, and just uses direct mappings.. :(
  • Is it larger than a factor of 2? I am willing to put up with a factor of 2 difference if it buys significant speed...

  • Your use of the word "kernel" in your response hints that I was unclear. I wasn't referring to Linux@64 bits, but rather the CPU. The comments about 64 bit programs being much larger than 32 bit ones suggests that there are different operating modes for the processor. These are the modes I was referring to.

    Sorry for being cryptic
  • I saw a lot of businesses using it for engineering purposes (CAD type stuff) on *very* high end hardware (RAID 5 with Cheetah drives, 3d mice, blah blah blah). $40,000 workstations. Just too expensive to get that much of a market share.

  • I'm surprised Cutler hasn't thrown a tantrum and left already...

    I'm surprised that Cutler can look himself in the face long enough to shave every day. VMS was (and still is, AFAIK) renowned around the world for its absolutely unquestioned rock-solid stability and bank-vault security. NT, his current love-child, is renowned for a lot of things, but stability and security aren't two of them. In his shoes, I would be consumed with shame.

    I'm guessing that he lays the blame firmly on Micros~1 management. Typical refuge of a failed technologist.

    Cutler's brilliant, but history will probably remember him first for NT, and only second for VMS. Sad, really.

  • You believe he was on the panel?

    Ummm... He said he was there. True, he could be lying, but it's probably in bad form to outright accuse him of such based solely on the fact that You don't know of an authoratative source. ;-)
  • Uh, you right, me wrong.

    Guess I took's comment out of context.


  • If MS wants their OS to run on a specific hardware platform, then THEY ought to do the f-ing port job. Just like any other OS vendor. NOT the hardware manufacturer.
    Did intel write the x86 port of BeOS?
    Did intel write the x86 version of Solaris?

    f-ing Microsoft needs to get off their high-horse, because that model was bullshit anyway.

    "The number of suckers born each minute doubles every 18 months."
  • NT supposedly has a POSIX compliance subsystem too.

    Supposedly is the key word there. It is almost completely unusable, and looks to be that way by design. It is often speculated that the only reason it exists at all is to provide a loophole to allow bids on government contracts that require POSIX compliance.

  • Because when push comes to shove, if Compaq wanted to crank out 200,000 desktop machines next quarter, they'd have no problem obtaining 200,000 intel CPUs and motherboards, but no hope of finding that many Alpha CPUs and motherboards.

    "The number of suckers born each minute doubles every 18 months."
  • Well, there was a Windows 286 (I still have the box around somewhere), but perhaps you typo'd that for Windows 386. And I'm not sure whether Win95B wasn't just another name for Win95 OSR2. (OBTW, there's also a Win 95C, which added some USB support to Win 95B).
    Then there are all the beta versions floating around. (I ran W95 beta for about six months after W95 was released, I couln't be bothered doing the upgrade, and besides I was usually booting to Linux instead.)

    Then of course there is (or I should say, was) Win NT for Alpha, Win NT for MIPS, Win NT for PowerPC, ...
  • When I got my first C compiler for DOS I bought it used without (as far as I could tell) a license. I called up the company, after waiting jut a few seconds I was online with a real tech, who offered to send me (free of charge) the latest disks... I swore then that this was the company for me....

    Today I curse that company's name... Microsoft.
  • What'd UNIX ever doto him? :)

    It killed VMS. :-)

  • Compaq, realising that it can not market Alpha servers to the ignorant Great Unwashed People in Suits, decides to give up on Alpha. Everyone wants NT anyway, and they have agood business selling x86 machines, IA64 is coming out sometime, even if Alpha is 5+ years ahead but they don't have to pay for Merced's development, so they (Compaq) pull support for developing NT on Alpha and quietly junk the architecture.

    That makes good business sense in the short term for Compaq, and Microsoft too because it's one less processor to develop for. If there's only one architecture in the market place (Intel) who needs cross-platform portability (in the eyes of the Great Unwashed)?

    Never mind, at least Sun are still plodding on with their stuff...

    You mark my words, in a few month's time you won't be able to buy a new Alpha for love nor money.
  • Well, shit yeah it's cheaper to develop Linux.

    MSVC + 1yr membership in MSDeveloperNet = $2000

    You do the math, but don't use a Pentium, because of the FP bug.

    "The number of suckers born each minute doubles every 18 months."
  • While I am not certain that the list is complete, this list [] of release dates for many MS operating systems can be handy to have around.

    One time it helped me was when someone claimed that NT 3.5 had been reported to have an uptime of 7 years. I could then point to this document that revealed that the version had only been out four or five years.

  • The person making the claim has to come up with the proof. If someone claims Bill said that, then they need to prove it.

    The default position is not to believe.
  • OK this is soo annoying! BEOS IS NOT BASED ON ANY UNIX LINUS BSD IRIX OR ANY OTHER UNIX CLONE. It is merely posix compatible (not completely compliant, however) meaning that it can run many unix applications. it however is in no way related to *unix. please tell me where you people are digging up this lie so I can tell the purveyer of this information to stop.
  • um. sorry to break the news to you, but JavaOS is dead.

    "The number of suckers born each minute doubles every 18 months."
  • That's now.

    Where's the future?

    "The number of suckers born each minute doubles every 18 months."
  • They already announced that the follow-up to Windows 98 will be called "Windows Millenium". Guts: DOS. GUI: Internet Explorer.
  • >The neat thing is: Linux has Merced support >*now*. No other OS can say that.

    Do not comment on that which you do not know.

    You have no idea how many in-house proprietary UNIX products are running on Merced "simulators" around the Valley.

    They just don't wear it on their shoulder like linux monkeys.

  • Intel does not own the Alpha exclusively.

    Compaq, Samsung and AMD all have access to the
    Alpha architecure and can more or less do what
    they want with it.
  • I don't know what the size difference is, but apparently it was enough that Linus figured that ia64 binaries would be slower than ia32 binaries(note: that ia64 and ia32 are completely different instruction sets, it's not just 64 vs 32 bits). We'll see when the chip is released. Also, I'm guessing here, but ia64 binaries will probally take a lot more time to compile because VLIW apparently requires a lot more compiler work.
  • I said I "messed with" the beta. I didn't "beta test", and I didn't have time to chase after something as glaringly obvious as this issue. But there is little chance of any kind of user-side resolution with error messaging as obtuse as this, so what would be the point?

    (If you're wondering how I got ahold of the beta if I'm not a "beta tester", it was provided with a win2k class I attended, still have it, still using it as a coaster.)

    "The number of suckers born each minute doubles every 18 months."
  • And how exactly working with NT makes one loose his integrity ?

    Those are computers we are taking about, computers...

    If you step back you will realizez than Linux is riding on hype more than on anything else ...
    At this point it is definately worse OS than Solaris and probably most of the commercial unices.

    I am not saying that Linux is worthless but you need to keep things in perspective ...

  • so many out here wanted one of those, and
    many did buy them, but many here were also stopped
    by the very expensive proprietary ram which you
    must use in them (and no, i have no reason to
    believe it any faster then PC-100 dimms, twice
    the bandwidth to the ram does not seem to make
    much difference) wether or not this is a case of
    it, sgi is known in the industry to find some way
    of getting more of you money. so some think this
    ram is just sgi back thier old tricks again.
  • NT supposedly has a POSIX compliance subsystem too.

    "The number of suckers born each minute doubles every 18 months."
  • But your list is deceptive. The service packs are generally just minor cosmetic or bug fixes and addition of user mode programs.. not kernel changes. Also Workstation and Server use identical kernels, the only difference between the two is only a single registry setting a a few extra user level programs.

    I must disagree. Some of the service packs for NT were significant enough changes to merit a version number change in any other sane numbering system.
  • I have news for you. You're average Unix Sys Admin makes alot more than your better than average MCSE

    Well, to be a Unix sysadmin, you actually have to KNOW something.

    The be an MSCE, you just have to pass a test. A test that is freely available on the net. Just write the answers on your hand.

  • Linux runs fine on a 386. It's not what you want for development; I do a lot of compile/test cycles, so I want more speed. (I know, I'm responding to a troll...). As for "B": it looks like those MSCE courses didn't teach you to read. "At least" means that what I was saying is that one GOOD thing (perhaps the ONLY good thing) about the MSCE program is keeping people off welfare.
  • In case you haven't noticed, the majority of Alpha sales have been OpenVMS and DG-UX kits, not NT. There actually hasn't been that much demand for NT on Alpha, so I think Compaq is just doing the savvy thing -- they're just dropping a less profitable (for them) OS.

    Dropping NT is not going to have that significant an effect on Alpha sales.
    Berlin-- []
  • Why not just use the Gimp and Blender or POV-Ray or BMRT or Radiance? Or a mix of all of them?
  • In the future, please refrain from rational postings in a "I make more money than you"/"My daddy can beat up your daddy" flame-threads. They exist for my amusement only, and your post might (off chance, I know) spur someone to reconsider posting another amusing (and probably false) "I make 50000000000 a second thanks to microsoft/redhat" post. And that would make me very angry...
  • Merced is a funny creature. It supports two different instruction sets. One is the x86 (or now ia32) instruction set in use on most PCs today. The other is the ia64 instruction set. ia64 is 64bit, while ia32 is 32bit, but there are a whole slew of other differences. ia64 is also VLIW. I have no clue what this means but apparently it requires quite a bit of compiler voodoo and will result in extremely large binaries. IIRC there is a mode switch involved to go from ia32 to ia64, but I don't remember how long it takes. It's probally very similar to switching from real mode to protected mode on an x86 cpu, which IIRC is very expensive.
  • Uuuhh... I wasn't talking about developers, my message was in response to one about Unix Sys Admins making a LOT more than the average MCSE.

    I know LOTS of MS developers, and yes, it's true, you don't need an MCSE to do MS development. In fact, I don't think I know of one MS developer who IS an MCSE (mainly Tech Support people get MCSEs anyway). But what you DO need to be an MS developer is an extraordinarily expensive copy of MSVC, and a subscription to MS Developer Net, which is outrageously expensive. Unless you Pir8 your SDKs. . .

    "The number of suckers born each minute doubles every 18 months."
  • Microsoft's policy appears to be that they will only support non-X86 versions of NT if someone else funds the port and the maintenance. That's what killed the earlier non-X86 ports. It seems that Compaq decided that they weren't selling enough Alpha NT boxes to justify the cost of keeping NT Alpha alive. Microsoft doesn't think that NT Alpha is important enough to pay for it out of their own pocket.
  • Yes, database servers.
  • by ToLu the Happy Furby ( 63586 ) on Thursday August 26, 1999 @02:38PM (#1724361)
    In the press on this topic, Compaq was quoted as saying that only 5% of their Alphas went out as NT systems.

    Actually, the figure I saw was 2%.

    Furthermore, MS is still using Alpha as the development system for 64-bit W2k; they just apparently won't be polishing it, marketing it, etc.

    Now, this seems pretty darn stupid of Compaq to me to discontinue plans to market 64-bit NT. Yeah, yeah, 2%--why bother, right? Thing is, that 2% refers to running a 32-bit OS on a 64-bit CPU. If you were going to the expense of buying an Alpha, why hinder it with an OS that doesn't take advantage of its architecture? Hence the 2% figure makes a lot of sense.

    But now that Intel is coming out with a 64-bit chip, well, Microsoft decides that's important enough to warrant a 64-bit OS...and this is the time Compaq chooses to pull the plug on Alpha NT?? Huh???

    Especially since, according to all reports, by the time Merced is finally released (Q2 next year? Q3??), the Alpha's going to be walloping it. By the middle of next year, the roadmaps I've seen have the 21264's hitting well over a GHz...I think I've even seen 1.4 tossed around (but it was prolly at The Register, so, grain of salt). More importantly, benchmark roadmaps (these, I remember, were leaked from Compaq, so more salt) have a 1GHz Alpha beating an 800MHz Merced by ~40-50% (SPECint, IIRC).

    So what's up???

    Two theories. Wait, make that three (er, four):

    1. Compaq's stated reason: they want to revitalize Tru64, and support Linux. Plus, no demand for NT Alpha. Possible...but I dunno. Tru64 seems to be a losing battle as a long term strategy; it's faster than Linux on Alphas now, but a few years down the road?? Seems smarter to go the sgi route, let your proprietary UNIX die, and let Linux sell your boxes. As for the lack of demand for 32-bit Alpha, see above. I think once it was 64-bit, people would buy. An appropriately high end Alpha box'd kick the hell out of an 8-way Xeon box on NT, and prolly a 16-Athlon as well.

    2. Merced NT doesn't have much in common with Alpha NT, so MS doesn't want to spend time on it. Here, I pretty much have little idea what I'm talking about, but it would make sense to me if programming for a VLIW chip is just too different from programming for anything else. Cause VLIW is waaay different--far more different from existing architectures than CISC is from RISC, at least the way "CISC" and "RISC" chips are made these days. So perhaps the whole, "we're making a 64-bit OS anyways, why not port to Alpha?" idea doesn't actually work. On the other hand, most VLIW issues seem to occur in the compiler...and they're still using Alpha as the development chip for 64-bit W2k. So...

    3. Maybe Merced'll be faster than we thought. This is my conspiracy theory of the day, and you have to understand that I've been pretty disappointed by Intel tech lately, so it's taking quite a bit for me to say this. But still...I mean, even if they didn't count on selling many Alphas for NT, don't you think Compaq would love to be able to claim how their much cheaper (and they will be) Alphas kicked the tar out of those new Merceds in a fair fight with the same OS?? (Note: yeah, it wouldn't actually be fair...but it'd sure look that way.) So either Compaq got wind of the fact that Merced won't be slower after all...or perhaps they realized MS wouldn't do what it takes to make the 64-bit Alpha W2k compete with the Merced W2k...or perhaps...

    4! Intel paid 'em off!! Ok, this is too fun to pass up. Essentially this goes like, Intel's just spent, what, 3 or 4 years now on a laughingstock of an overpriced chip (yes, I have no doubt that eventually IA64 will be amazing...but not on Merced), and they desperately want to avoid any controlled comparisons. So they politely ask Compaq and MS to stop development on 64-bit NT for Alpha. Besides, Compaq'll be selling their share of Merced boxes, so they've got as much to lose as anyone.

    Pretty spooky, huh!!
  • For example, most of the Win32 interfaces (as well as driver interfaces etc) are phrased in terms of 32-bit data objects (ok) and 32-bit pointers (oops!)

    So, do you shift all interfaces to 64-bits? Add a second set of 64-bit interfaces?!? Tough to add 64-bit support in NT, and still retain some useful degree of backwards-compatibility for 32-bit software.
  • One of us has misread the articles. From what I've read, Compaq will still happily sell you a fully support Alpha system.

    You just have to run Linux on it. Compaq simply performed a cost-benefit analysis on continued support for NT/Alpha under the current terms from Micros~1, and decided it wasn't worth it. (E.g., to cover their costs they would have to raise their prices to the point where people would buy Intel systems, and since that reduced the user base they would have to raise the prices even higher, in a viscious death spiral.)

    Had Micros~1 agreed to pick up more of the development and support costs, to develop the market, Compaq might have reversed itself. But Micros~1 has apparently decided to tie its fortune to the Pentium chip. (Not Intel, until they have a working 64-bit system. Pentium.)
  • Goody suggested:
    Get someone like the people who make VMWare to get Win9* to run on Alpha and start packaging RedHat on Alpha machines.

    VMWare? Why bother? Compaq now owns the original translate-on-the-fly Windows emulator, DEC's excellent FX!32 []. All others are pale imitations.

    The trouble is that DEC had a sever case of craniorectal insertion (I know, I worked with them on two horribly mismanaged Alpha projects.) To this day Compaq tries to charge more for DEC Unix than a high-end workstation hardware costs. They just don't grasp the importance of GIVING AWAY the FX!32 software as the only way to bootstrap their hardware sales.

    Tragedy. In the Greek sense.
  • Hasn't Linux been 64-bit ready for years now?
  • by jd ( 1658 ) <> on Thursday August 26, 1999 @04:05AM (#1724403) Homepage Journal
    Microsoft's software has enough trouble running in 32-bits. The changeover from 16-bits to 32 was positively painful for them, and that was with near-100% compatiable processors.

    The idea that they could take Windows 2000 (or any other version of Windows) and simply compile it to work on the 64-bit Alpha, or Merced processors was always going to be a joke.

  • Microsofts products are too fragmented and they don't have any development roadmap for the future :)
  • MS hasn't run Win64's source through their Code-Bloat Wizard 2000(tm) yet.
    It's that simple.

  • Wasn't it Intel that was buying all the Linux books? Maybe now we know why.

  • by Shoeboy ( 16224 ) on Thursday August 26, 1999 @04:10AM (#1724438) Homepage
    What happened to "NT is extremely portable." In theory, the only hardware specific component is supposed to be HAL.DLL. The rest of the NT code base is supposed to be highly portable C code. By replacing HAL.DLL and recompiling the rest of the system, it was supposed to be possible to port NT to whatever in a matter of weeks. Why did compaq need an NT development team in the first place? Why did NT powerPC collapse when IBM withdrew funding? I refuse to believe that MS doesn't know how to write architecture neutral code - that's just too far fetched. I'm thinking that it's due to lightning storms.
  • The way Microsoft phrase it, it's all Compaq's fault for "terminating Alpha support for current versions of Windows NT". As if Microsoft were helpless in front of Compaq's refusal to let them port their operating system!

    Linux, as everyone here knows, ran on Alpha machines (as well as x86 and 68000) long before it could get acknowledgement of its existence from the processor manufacturers, let alone substantial help in making it happen. It was done in the usual way: get the specs of the hardware, and code to them. What is it that's so difficult about this process that Microsoft needs Compaq to hold its hand before it can think about it?

    Microsoft don't expect ever to have to act like a software firm. You don't write code for other people's hardware - you graciously allow the hardware manufacturer to write code for you. Compaq have started to smell the independence from this kind of treatment that Linux gives them, and soon Microsoft will find out what that means for them.

    Revenge is gonna be so sweet.
  • According to Microsoft's site:
    On August 23, 1999 Compaq decided it would terminate Alpha support for the current versions of the Windows NT operating systems... Compaq also decided that it does not plan to support future Alpha versions of Windows® and BackOffice® including Windows 2000.
    According to the Netcenter article:
    The controversy over commitment to Alpha started late last week when Compaq informed about 100 engineers at its Bellevue, Washington, facility of impending layoffs. Many of the engineers were involved in the development of Windows NT and Windows 2000 on Alpha... "Compaq failed to communicate clearly within, with Microsoft, industry analysts, customers, and the media," said one source, who asked not to be identified
    I dunno, this sounds more like it signifies big trouble for Compaq than Microsoft.

    As for the Merced announcement, I have to plead ignorance. Is this really late enough in the game to signify big trouble for MS? Or is it a normal kind of problem we should be expecting here? (I guess I'm assuming the two aren't mutually exclusive...)

  • by LizardKing ( 5245 ) on Thursday August 26, 1999 @04:23AM (#1724468)
    The NT concept of a hardware abstraction layer that marshalled calls to the underlying system was fundamental in making NT stable. This is why NT3.51 was very stable. However, it also made many things slow, most noticably the grraphics side of things. With NT the GUI is not optional, so this poor performance was always noticable. To improve performance, graphics operations were allowed to bypass the HAL in NT4.0. This is why NT4.0 outperforms NT3.51, but is very unstable.

    I don't know whether any other userland stuff was allowed direct hardware access, as I quit using NT after 3.51.

    Chris Wareham
  • Remember Access 1.0? Remember Windows 1.0?
    Products that claimed to work - that sometimes worked - just enough to get a few people to buy into the concept, and fund development of another version. Granted, Access 97 and Windows 4.0 work /*much better than their predecessors*/, so the strategy worked. But I guess you just can't pull that crap with a whole new arch. Especially with Linux on the horizon - it's nice to know that 'the people have voted', and 64 bits on the desktop will be a reality, sans M$.

    Now, about that 'chasing tail-lights' thing...
  • This document somehow leaked out of compaq - darn pesky email ... enjoy the "official" compaq position .. indaba...

    Here is part of the communication package around Alpha NT.

    No more NT4 on Alpha past SP6.
    No 32 bit Win2000 on Alpha.
    No 64 bit Win2000 on Alpha.

    Alpha is purely for Tru64 Unix, OpenVMS and Linux.

    Windows NT Alpha Strategy Q & A Version 1 - 25 Aug 99

    General Questions:

    Q. Exactly what has Compaq changed in its Alpha Windows NT Strategy? What has Compaq announced? Why has Compaq decommitted from Windows NT on Alpha systems?

    A. Based on the extremely good scalability of our 4- and 8-way ProLiant servers with 32-bit Windows NT, Compaq believes that this platform can satisfy all market requirements for 32-bit Windows NT. We have therefore been able to simplify our strategy and offerings. Compaq will end 32-bit Windows NT Alpha systems development with V4 SP6, late in 1999, and will not support either 32 or 64-bit Windows 2000 on Alpha systems.

    This change will enable us to sharply focus our Alpha strategy and resources on our aggressive plans to grow Tru64 UNIX market share, support our loyal OpenVMS customers, extend our Himalaya range, and drive volumes for Alpha systems with Linux.

    Q. What is Compaq's Windows NT strategy going forward?

    A. As the leading provider of Windows NT-based platforms and solutions, Compaq is a strong supporter of Windows NT and will remain at the forefront of moving Windows NT into the enterprise as its capabilities continue to mature. We will maintain our leadership role in providing the environment in which Microsoft is developing its 64-bit capable versions of NT as well as our extensive involvement in assuring the best performance and reliability on Microsoft's current and future 32 bit offerings. The only change in strategy is that all of Compaq's efforts on behalf of 32-bit Windows NT will now be built around our industry-leading IA32-based systems. The recent very strong performance results achieved with Compaq's new 8-way ProLiant servers demonstrates that we can meet 100% of the market requirement for 32-bit Windows NT systems with these platforms.

    Q. Will Alpha support for 64-bit Windows NT also be discontinued? Why?

    A. We will continue partnering aggressively with Microsoft on development of 64-bit Windows NT, utilizing Alpha systems. We do not plan to offer 64-bit Windows 2000 on Alpha systems, and will focus our efforts on offering the very best 64-bit IA32 Windows NT platforms in the market at the time of its introduction.

    Q. Is this a sign of Compaq slowly backing away from Alpha?

    A. Absolutely not. Alpha still remains a vital component of Compaq's NonStop eBusiness strategy for the enterprise. As evidence of this, the previous commitment to move the NonStop Himalaya and Integrity system architectures to Alpha is intact, as is the recent commitment by the senior management to spend an incremental $100M to further the position of Tru64 UNIX/AlphaServer in the marketplace. There is also strong market interest in Linux running on Alpha, and this will be a major focus for driving volume based on the Alpha architecture. Finally, the Alpha chip roadmap continues advancing, with "shrinks" of the 3rd-generation of Alpha architecture (the 21264), the finalization of the EV7 design, and early design of EV8 technologies all presently underway.

    Installed Base Impact Questions:

    Q. Am I "dead-ended" with the current investment in my AlphaServer? What options do I have?

    A. No. Compaq is putting in place programs and growth paths to satisfy our customers' requirements through this change in plans.

    For already installed AlphaServers running Windows NT: In the short term, customers can continue to use their existing systems with current applications and have the option to upgrade to Service Pack 5 and then Service Pack 6. In the long term, customers can continue to use their current systems with Windows NT 4, and will be supported by Compaq Customer Services until at least Q1 CY2001. For future needs (including Windows 2000), customers should take advantage of trade-in programs and migration services to Windows NT or Windows 2000 on ProLiant servers, or to Tru64 UNIX, OpenVMS or Linux on their AlphaServer systems.

    For customers considering new AlphaServer systems running Windows NT: Customers can move to ProLiant Servers and run Windows NT 4 or Windows 2000, depending on their requirements.

    For each of these circumstances, Compaq will provide attractive programs, upgrades and migration services to fit the individual customer needs. Details on these programs are available from your regional representative.

    Q. I bought my Tru64 UNIX or OpenVMS AlphaServer with the promise that it would run Windows NT in the future. What do I do now?

    A. Compaq will provide upgrade and trade-in programs to ProLiant Servers for customers who chose to move to Windows NT in the future. Details of these programs are available from your regional representative.

    Q. I have an Alpha Windows NT-only DIGITAL Servers (white box), what options do I have?

    A. Customers with Alpha Windows NT-only DIGITAL Servers can continue to run Windows NT V4 with support provided up through at least Service Pack 5. For customers who wish to move to Windows 2000, Compaq will provide attractive upgrade and trade-in programs to new ProLiant Servers.

    Q. How long will my current Windows NT Alpha system be supported?

    A. Our Customer Services organization will continue to support both hardware and software, for the foreseeable future and at least through Q1, 2001. Hardware support will be offered under standard terms & conditions. As a demonstration of this commitment, Compaq Customer Services still supports VAX and PDP technology.

    Q. How long will you ship new Alpha systems running Windows NT V4?

    A. A specific roadmap that details the last ship dates for each system model is provided in supporting materials.

    Q. I have a large investment in Alpha for UNIX, why won't Tru64 UNIX be next to be dropped (when Merced ships)?

    A. The market for enterprise servers running UNIX remains a very strong and vital one, and Compaq intends to be a leader within the segments of this market we focus on, in order to service the needs of our enterprise customers. Tru64 UNIX is widely acknowledged to be one of the finest UNIX operating systems in the market. The recent commitment by senior management to spend an incremental $100 million to fortify the position of Tru64 UNIX/AlphaServers in the market is tangible evidence of our intentions.

    Q. I am in the middle of Y2K validation and am planning to move to Windows 2000. Why has Compaq dropped Alpha support for Windows NT at this critical time?

    A. Customers will continue to be able to run Windows NT Service Pack 4 for Y2K conformance and have the ability to upgrade through Service Pack 6. This will enable customers to run current Alpha systems through CY2000 without impacting their Y2K readiness. As customers begin to move to implement Windows 2000, they have the option to re-deploy their current AlphaServer to run Tru64 UNIX or OpenVMS, add ProLiant Servers running Windows 2000 or take advantage of upgrade and trade-in programs being put in place by Compaq.

    Q: How will this decision affect StorageWorks? Will Compaq StorageWorks products continue to support NT on Alpha?

    A: The Raid Array 8000 and Enterprise Storage Array 12000 product lines will support 32-bit Windows NT on AlphaServers in FCAL configurations. Customers will have the option to upgrade to Service Pack 5 and then Service Pack 6. Compaq is making beta versions of the FCAL drivers available today and will have a production FCAL driver available for AlphaServers 32-bit Windows NT in October, 1999. This solution will support both clusters and high availability single system configurations. In the long term, customers can continue to use their current systems with Windows NT 4, and will be supported by Compaq Customer Services until at least Q1 CY2001. For future needs (including Windows 2000, support for FC Switched Fabrics, and value-added features such as Data Replication Manager), they can move to ProLiant Servers or consider trade-in programs and migration services to Tru64 UNIX, OpenVMS or Linux.

    Q. What will happen with my ISV software licenses as I move to another platform?

    We are working closely with key ISVs, for example Oracle and SAP, to assure that there are migration plans in place. Our key ISVs are supportive of this decision and are also committed to taking care of customers in the very best way. More details will be provided as each ISV develops their plans.

    Business Questions:

    Q. Was Compaq counting on Alpha Windows NT to drive volume?

    A. No. There is enough Tru64 UNIX, OpenVMS, Himalaya and Linux business to sustain a competitive Alpha processor for the long term, along with our partners such as Samsung/API.

    Q. What is Microsoft's position on this decision? Does this reflect a weakening of the Compaq/Microsoft relationship?

    A. Microsoft supports this decision. It in no way reflects a weakening of the Compaq/Microsoft relationship. Compaq is working quite closely with Microsoft in the development of 64-bit Windows NT, in support of our drive to address more demanding enterprise requirements with Windows NT.

    Q. How have Samsung and Alpha Processors, Inc. reacted to this decision? What will happen to other OEMs who are developing systems utilizing Windows NT on Alpha?

    A. Recognizing the opportunity that exists for Alpha on Linux, Alpha Processor, Inc. has already made all its products available for Linux. Most recently they've been assisting an effort in bringing leading Linux tools and libraries to Alpha. API's and Samsung's resources will continue to support Linux-based solutions. With respect to OEM's who have been utilizing Windows NT on Alpha, API and Samsung are working closely to help those who wish to transition to other platforms. We'd recommend contacting API for further discussion regarding their platform strategies.

    Q. What proportion of AlphaServer business is on Windows NT today? Has that fraction been increasing or decreasing?

    A. Less than 2% of AlphaServer current business. Tru64 UNIX has become the predominant choice for Alpha systems, with OpenVMS in second position. While Tru64 UNIX and OpenVMS customers have benefited from Alpha's accelerated performance with EV6 technology, 32-bit Windows NT customers needs can be met with the performance and capabilities of the superior performing ProLiant servers.

    Q. What proportion of AlphaServer business is on Linux today?

    A. Small, but growing very rapidly.
    Q. What has been the reaction of ISV's? What about those ISV's who have made significant investments in Alpha/NT?

    A. Our key ISVs support this decision. We will be working closely with our ISVs to offer trade-in and migration programs, to take care our mutual customers needs. In addition, Microsoft has committed to support ISVs with all aspects of their 64-bit development program.

    Q. What is the impact of this decision on Compaq's workstation business?

    A. The decision regarding Windows NT and Windows 2000 affects both AlphaServer systems as well as Alpha-based workstations. Attractive migration and trade-in offers will also be in place to IA32-based Compaq Professional Workstations or to other operating systems on the current platform.

    Q. How will this announcement impact Compaq's focus on the ISP market?

    A. It does not affect our comprehensive and aggressive focus on the ISP market, addressed with a combination of Tru64 UNIX on AlphaServers and Windows NT on ProLiant servers.

    Overall Server and Competitive Positioning Questions

    Q. How are Tru64 UNIX AlphaServer solutions for ISPs, ASPs and eBusiness positioned in the market?
    A. Compaq Tru64 UNIX AlphaServer solutions are competitively positioned. Unlike any other RISC UNIX platform on the market, solutions delivered on Tru64 UNIX V5.0 AlphaServer systems offer our customers the highest availability with the lowest cost to implement, operate and manage. Additionally, eBusiness applications require the ability to scale to meet the peak demands placed on them. Unlike other UNIX platforms, Compaq's Tru64 UNIX AlphaServers have the power to meet the peak load requirements. Your systems will not slow down during a critical time of network traffic.
    Q. Compaq appears to be positioning the ProLiant 8-way server against Sun. Isn't that in conflict with the positioning of Tru64 UNIX AlphaServer solutions?
    A. Not at all. Compaq has a stronger set of products than Sun since we have both superior UNIX and Windows NT products, offering the customer more choice! The newly announced ProLiant 8000 and 8500 servers were announced to compete directly against Sun in the ASP and low-end of ISP markets, complemented by Tru64 UNIX for the higher end requirements. Compaq Tru64 UNIX AlphaServer solutions continue to compete effectively against Sun as demonstrated by numerous industry standard benchmarks and with the extremely high levels of availability and scalability.
    Q: What are the benchmarks and how do Compaq ProLiant and AlphaServer offerings compare with Sun?
    A. As an example, Compaq's AlphaServer DS10 with Tru64 UNIX, featuring 2GB memory and up to 54 GB disks, outperformed Sun's dual- and quad-processor servers running Internet applications at less than half the price on the SPECweb96 benchmark. The AlphaServer DS10 with Tru64 UNIX delivered results of 3404, outperforming both the Sun E450(2963) and the Sun E250 (2625). There are many other examples in other target market segments.
    The ProLiant 8000 running Microsoft SQL Server 7.0 on Windows NT 4.0 offered more than double the performance at 30 percent better price:performance than the Sun E450.
    Our key strength versus Sun is that no matter which OS our customers chose, either Tru64 UNIX on AlphaServers or Windows NT on ProLiant servers, they can be assured of the best absolute performance as well as price/performance across a range of applications and systems. Sun only has UNIX and cannot offer their customers the Windows NT choice.
    Q. Isn't Compaq's strong support of both Windows NT and UNIX solutions confusing to the market?
    A. Compaq is committed to solving the full range of business and technical challenges throughout the enterprise. Analyst reports estimate that 90% of all IT environments will be using both UNIX and Windows NT. Compaq, a leader in 64-bit UNIX and the market share leader in Windows NT solutions, is not only committed to offering a broad range of solutions for each operating system, but also to continuing to enhance its leadership position in UNIX/Windows NT interoperability. Compaq allows its customers to take advantage of both systems - without two support and development staffs, and a host of cross-platform application issues.

    Workstation Alpha NT Strategy

    Q: Why did Compaq decide to discontinue Windows NT on the Alpha workstation platforms?

    A: Compaq re-evaluated the Windows NT strategy for workstations because of vastly improved graphics capabilities on Compaq AP and SP Professional Workstations and limited Windows NT sales on the Alpha workstations.

    Q: How does this announcement effect Alpha Windows NT customers?

    A: Compaq continues to ship and support Windows NT version 4.0 (Service Pack 3) on current Alpha workstation platforms (DPW500a, 600a, and XP1000/500MHz).

    Compaq will qualify, release, and support Microsoft Service Pack 5 and Service Pack 6 for current Alpha workstation products (DPW500a, 600a, and XP1000/500MHz).

    Q: How will this announcement effect Alpha workstation products?

    A: Compaq will not support Windows NT or Windows 2000 on future Alpha products.

    Platform upgrades and future Alpha workstations will support Tru64 UNIX, OpenVMS, and certified Linux operating systems.

    Q: Will Compaq support 64-bit Windows NT on the Alpha workstation platforms?

    A: There are no plans to deploy or support 64-bit Windows NT on Alpha workstations.

    Q: How does this announcement effect Tru64 UNIX, OpenVMS, and Linux Alpha customers?

    A: This announcement does not directly affect Tru64 UNIX, OpenVMS, or Linux Alpha workstation customers. We expect to re-deploy certain resources from Windows NT to these operating systems in order to provide a higher level of support there.

    Q: How does this announcement effect Compaq Intel workstations?

    A: This announcement does not affect Compaq AP and SP Professional Workstation customers.

    Q: What is the migration path for Windows NT Alpha workstation customers?

    A: Compaq will work with customers to help them migrate to the optimum AP or SP Professional Workstation or to one of the other operating systems supported on Alpha.

    Q: What does this decision mean for Windows NT Alpha workstation OEM customers?

    A: Compaq will work with OEM customers on an individual basis to determine how best to meet their short and long term needs. This can include OEM specific last buy dates and technical support for hardware and/or software migration plans.
  • I've got a ruffian board, 512Mb RAM (couldn't get the 6 SDRams to work with 2.0.36) and a 600MHz alpha.... on the good side, I've got a 3C905B to work instead of the onboard tulip and same for SCSI, I use an adaptec 2940UW but have to really check what difference it makes on my transfer rate, because now I have to boot from a floppy! (milo won't boot from this 3Mb /dev/sda1 partition)

    No glitches in Debian when you get round the really confusing documentation, and pretty much everything I use compiles great and seem to work fine. WM, Gimp, PDL, LaTeX...

    So why am I complaining? Well, if you try some of the logo scripts-fu on both the alpha and a PIII-550, the speed difference is not that noticeable at all.

    The real test would be to do a povbench on both machines... but I got some errors half way thru the rendering on the alpha!

    When alphalinux gets a descent compiler (I'm just blaming compaq here, the gcc team is doing a great job no matter what!)... well, that's going to be a different story, but compaq will never allow Linux take shares of Tru64 or whatever the name is... (OSF)

    I'll check bladeenc with compaq's enhanced mathlib... that should be a pretty good test, but until then, I'm not convinced.


  • Ummm, you've obviously never heard of Windows 1.0. Oh, wait, it was just plain Windows back then, wasn't it. It was character based and really didn't work well. Then there was Windows 2.0. That didn't work well either, but at least it task-switched, and the mouse worked, though the windows could only tile at that point, sort of like dosshell. Then came Windows 3.0, and the rest is, as they say, hysteria.

    Win4.0 was going to be the market name of what became Windows NT, er, Win2k, er... (Linux3000 - Imagine that! Software available 1000 years ahead of schedule!) If I recall the marketting FUD correctly, this was codenamed CAIRO. Win95 was MEMPHIS, and WFW was DAYTONA, isn't that right? I ask, since you're apparently in-the-know of things MICROS~1.

    As for the evolutionary software development process - I'm all for it. But charging customers full-price for a job half-done is unethical. Advertising non-existent or inadequately tested features, selling vapor, and charging for bug-fixes is wrong. Committing to support a product, and bailing out when the company with actual know-how does, is BS. This is what seems to have happened with the Alpha.

    This is what Microsoft tends to do each time they venture into new territory. A new product line always debuts as a dud, albeit an expensive and well-marketted one that funds the development of a v2.0... It's enough to make the sale, and then once you've got'em, charge'em for the upgrade (at a lower cost - so it looks like you're doing them a favor - Puh-Leez!)

    Now, don't get me too wrong. I use NT at home and at work. Nice and solid - no problems in over four years... (on 32-bit Intel) But MS seems to care only about the upgrade cycle and time-to-market, not about true quality and customer satisfaction. If the originator of the idea isn't there to help them, they fold, or just pile on useless features. This is why they're starting to slip. Innovation indeed.

    The thing with Linux is, as I'm sure we all know by now, that it's free. If it's got a bug, it get fixed quickly, and for free. If it isn't getting fixed, you're free to do it yourself. The only free solution to Windows problems is Linux. Funny that.
  • Compaq had said that they would support NT for Alpha64. This sounds like MS is saying, "Oh yeah? Drop Alpha32 and kiss Alpha64 goodbye!".

    It's interesting to see this in the light of yesterday's "Interview with Original NT OS/2" developers, which stressed the importance of platform independence and portability. With this announcement (correct me if I'm wrong), MS has finally gone from 4 platforms (PPC/MIPS/Alpha/x86) to 1 (x86).
  • by dirty ( 13560 ) <> on Thursday August 26, 1999 @04:34AM (#1724507)
    Yes, on alpha it's been fully 64bit for a long time now. On UltraSPARC the kernel runs at 64bit but the apps still run at 32bit. So the support for 64bit has been there for a long time. I'm assuming the ia64 port is 64bit atleast for the kernel. I remeber reading something from Linus a while ago where he was saying that because the ia64 binaries are probally going to be so large it would be faster to run 32bit x86 binaries most of the time than to run ia64 binaries. The exception being binaries that need access to huge ammounts of memory, such as large databases.
  • You mean "Trade in your new Huffy for a Ross"?

  • No irony at all actually. NT is still highly portable. It runs on 486, pentium, pentium pro, pentium II and K-6 chips. What more portability could you ask for ;-)
  • by Goody ( 23843 ) on Thursday August 26, 1999 @04:41AM (#1724533) Journal
    What I can't understand is why Compaq can't get behind the Alpha and drive it home. It's superior over Intel architecture. Compaq could say "Screw you, M$ AND Intel" and have Alphas on every desktop. Get someone like the people who make VMWare to get Win9* to run on Alpha and start packaging RedHat on Alpha machines. Hell, they bought Digital and Tandem. If they knew how to integrate this vast wealth of knowledge together, they could kick ass, but apparently someone at Compaq just doesn't want to do it.

    M$ will get NT working on 64 bits, but obviously they've got egg on their face. So much for portability. This is certainly going to drain resources. How many versions of Windows will M$ be supporting now ? :-)

  • It looks like they are heading on the out. I think that they just pissed off one to many companies, and now these compaines are seeing Linux as a way to circumvent Microsoft.

    Rather than rewrite their OS, they can use Linux on there hardware. (Linux runs on Mac, Alphas, INtel, Sparc, etc. It make more sense to have one operating system that runs on many cpu's than many operating systems that run on many cpu's.

    This is sort of the Java thinking. Java runs on may Operating Systems, but it is a layer above the Operating System. This makes porting form one cpu to another easier.

    Most Linux software on Linux/Intel will run on Linux/Alpha, with just a recompile, or on Linux PPC. And it will run faster than Java, unless you compile Java to OS native code.

    How will Linux eventually affect Java development? In 5 years some Linux hybred OS may be the new OS. But a hybred of Linux and what? (so far SGI is adding to the Linux melting pot).

  • The inevitable decline of the Great Beast has indeed begun. This is a strong signal. Microsoft is realizing that with the very real threat of Linux, they can no longer play around with unimportant (as Microsoft sees them) platforms such as Alpha or PPC. They must concentrate on their core products/platform.

    BTW, what happened to all the Astroturfers that were posting yesterday about how the Compaq decision was of little importance, because 64-bit NT Alpha development was going strong?

    Whaddya think, everyone? Exit Cutler? Or is Merced going to keep him happy?

    Interested in XFMail? New XFMail home page []

  • by RedGuard ( 16401 ) on Thursday August 26, 1999 @04:48AM (#1724550)
    Strictly speaking nothing in user-land is allowed
    access to the hardware in any version of nt,
    but after version 3.51 the code that implements
    the gui was moved into kernel mode. In any
    version of nt sections of the graphics drivers lived in kernel mode and had access to the hardware, what was apparently slowed nt 3.5/3.51
    was the code path for gui calls which had to
    go to another process, in nt4 they go directly to
    the kernel.
  • by Cato ( 8296 ) on Thursday August 26, 1999 @04:49AM (#1724552)
    I'm no NT expert, but this is confused.

    What happened is that up to NT 3.51 inclusive, the Win32 subsystem was a completely separate process, as were the OS/2 and POSIX subsystems. In NT4, the Win32 subsystem was migrated into the kernel (in Linux speak; in NT it's the Executive, the Kernel is a very low level layer just above the HAL). There are rumours that this was resisted by key people on the NT team, but there was a strong push to improve GUI performance; there are also rumours that this is the reason for NT4 being less stable than NT 3.5x.

    Certainly I never had a crash on NT 3.5x and have had many on NT4, but YMMV. On some configurations, NT does not really crash much (e.g. my Linux workstation on top of VMware!).

    As for the HAL originally hiding all hardware - this is not true, otherwise why would device drivers be necessary? I think the HAL abstracts basic resources such as CPU, memory, bus, etc, but many other resources require drivers.

    I'm not hugely surprised by Linux doing better on IA-64, since Linux has been 64-bit for some time and has worked out a lot of generic 64-bitness bugs, whereas Win64 is still alpha.
  • Actually, Jim Paradis at Digital was pivotal in getting the Alpha division interested in Linux, and has since been important in speeding along Linux Alpha's success.

    Just some credit where it's due...
  • This is completely false. Very little real software ports that easily unless it was designed on a 64 bit architecture. Too many people think that sizeof(int) == sizeof (char *) which is not true on alpha.

    If your design and code is already 64 bit clean and abstract, you'll be fine. If not, have fun chasing down memory errors from assigning pointers to ints.
  • by CrosseyedPainless ( 27978 ) on Thursday August 26, 1999 @05:01AM (#1724576) Homepage
    And what would be thy problem with the horse-drawn cart, Englishman? It is not a tool of the Devil, like the forsaken-by-God products of Microsoft!

    dictated to me by an Amish moderator
  • As long as he wants to "kill Unix", I don't think there's any other company out there who needs him. The only non-Unix operating system I know of in development is Be, and I don't think he'd like the Gassee co-existance philosophy. After reading 'Showstopper', I can almost hear him scream "Coexistance is for wimps!"

    Of course the good news is that he's pretty rich now from Microsoft stock options, so he could always retire. But what then? I don't see him sitting around playing golf.


  • I'd have to agree. While I won't hold back criticizing the MSFT when they deserve it, I'm not sure if this is their problem. In other words, Microsoft seems to want to support the Alpha, at least a little, but Compaq is blowing away the NT Unit because it's "underperforming" for them.

    As for the Merced thing, I'd have to agree. This sounds more like an early hiccup vs. any serious problem. Believe me when I say that Intel will not hold back any support for getting MSFT stuff to run on their processors. Too big a market.

    Although, if Intel were to not help MSFT, it's be interesting to see who that would hurt more? Intel or MSFT? I mean, if you think about it, with this move, NT is basically a one architecture beast. So unless they start pushing AMD K7's as their premier processor (not a bad idea from the reviews of the processor), MSFT is really in some trouble. On the other hand, cheap Intel processors would probably aid Be, Linux, OS/2, BSD, and the number of other x86 operating systems out there.... be interesting, if you ask me.


  • by Ami Ganguli ( 921 ) on Thursday August 26, 1999 @08:57AM (#1724614) Homepage

    1. Compaq is probably not selling enough Tru64 or Linux systems to justify continued Alpha production. If Compaq was making lots of money, they might make some sort of daring move to push Unix/Linux on their systems. But since they're trying to cut costs they'll have to drop the Alpha.

    2. Microsoft could easily continue Alpha development. Windows on Alpha is of strategic importance when dealing with Intel, so MS is shooting themselves in the foot by letting it die.

    3. Microsoft isn't stupid. They're quite aware of points 1 & 2.

    Conclusion: Compaq must have done something to really piss off Microsoft. Maybe those NT engineers Compaq layed off were really important to Microsoft for some reason. Perhaps Compaq is about to make some sort of big Linux announcement. Maybe Compaq pulled out of a product tying deal. Who knows. Whatever it was, Microsoft now feels that it must retaliate against Compaq even though it will hurt Microsoft on another front.

  • Regarding this article: 99Aug04.asp (posted on Aug 25 on The original developers are bragging about how good their design and plan was (for NT).. They say: "It cost us a lot to keep portability alive, but we did, and that has made it easy for us to respond to things like Merced," he says, referring to the 64-bit chip from Intel. Well obviously they didn't make it that easy.. ;)
  • by mattdm ( 1931 ) on Thursday August 26, 1999 @06:22AM (#1724620) Homepage
    Have you used Linux on an Alpha? I have, and I didn't find it so superior. Last I saw, kernel modules weren't working, for example. Maybe that's changed, but still....

    Also, VMware isn't going to help Win95 run on alpha -- it relies on having a native CPU.


  • by Anonymous Coward
    100 programmers costs in excess of $10 million a year. This only made sense to Compaq if they got a good cut and sell many copies.

    Now, Microsoft being what they are, I'd bet Compaq knows they're indirectly funding Microsoft's entry into IA-64. Compaq pays the freight to pave the way from 32 to 64 bits. Then Microsoft does their own IA-64 port, cuts Compaq out of the revenue stream, and undermines their Alpha market.

    Standard Operating Proceedure.

    I'm glad to see someone wising up to Microsoft's business practices enough to make rational decisions. Microsoft will play along until its time for someone to get rich, then the one that gets rich will be Microsoft.

  • by SoftwareJanitor ( 15983 ) on Thursday August 26, 1999 @09:08AM (#1724626)
    Yeah, they are falling apart:

    That isn't what I said, exactly. While Microsoft is rolling in cash, it is primarily due to their control of OS preloads of Windows 98 and sales of Office. NT isn't responsible for their huge profits because it not only hasn't sold in huge quantities, it has cost them huge amounts in development and advertising. NT has failed to light a fire under the industry. It slowed, but failed to kill Novell. Novell looks resurgeant. It has failed to kill UNIX. UNIX growth continues. It has failed to slow acceptance of Linux. Linux is growing at a much faster rate than NT, despite the huge inequity of resources spent on promotion of NT versus that spent on promotion of Linux. Heck, NT has failed to even completely kill oS/2, despite IBM's best efforts at helping them out there.

    Lets get some realism here....

    The reality is that a lot of once large and once prosperous companies that relied on their own proprietary OSes have seen those products dry up and either ended up dead (like Prime (PrimeOS) for example), becoming a UNIX vendor (like HP (MPE), DEC (VMS), etc) or basically exiting the hardware/OS business altogether (Unisys).

    Its easy to predict the status quo when a company is prosperous, but that can turn around quickly. If you had told people in 1980 that Digital Research would be a tiny, forgotten subsidiary of a company whos major product was a distribution of a free UNIX clone, they would have told you that you were insane.

"I don't believe in sweeping social change being manifested by one person, unless he has an atomic weapon." -- Howard Chaykin