Sandy Bridge + Light Peak = EOL for Mac Pro & Xserve

Discussion in 'Mac Pro' started by fw3857, Jan 10, 2010.

  1. fw3857 macrumors newbie

    Jun 17, 2004
    After observing him for 25 years, I know how Steve Jobs' mind works.

    2011 Sandy Bridge = more than enough power for any user Apple cares about. [notice I didn't say enough power for every user that would like to use a Mac] Six cores in every remaining product will be enough for everyone by decree. OpenCL will take up some slack here at well.

    2011 Light Peak = the permanent end of slots. Remember how Steve fought against "slots" for the Apple II? Remember how the original Macs had none? Remember how he loved the Cube? How he loves the Air? Fundamentally, Steve is a minimalist, and truly hates anything extraneous. And by extraneous, I mean to him, not to you.

    I do believe that all Macs will only ship with audio ports, and Light Peaks ports. No FireWire, no USB, no ethernet, no display outputs.

    But what about all my legacy stuff?? Too bad. That is why there is eBay. Apple has never hesitated to drop old formats and connectors, and has never been shy in doing this.

    With the intro of the Mac Mini Server, the Xserve is for sure dead, maybe this year, or for sure the next. Apple cares nothing for the HPC space (no Infiniband drivers even). And Macs will never be used in the data center in any meaningful way for a huge number of reasons. I have no doubt that linux will be running on cheap white boxes at Apple's new place in NC. I'm really surprised that Xgird, Xsan, and so on are even still offered. Anybody hoping for mac "enterprise" products of any kind from Steve, or from any of his friends (Ellison, etc.) are SOL. Apple is all about the SoHo space.

    Pro Audio:
    No doubt some Pro Tools guys are in denial, but Steve has declared total war on Avid/Digidesign with their Apogee/Euphonix/Logic alliance. And by the ridiculously cheap pricing of FCP. Further, it would take nothing for Apogee to design Light Peak X-Symphony cards, and the dsp guys (UA, TC, etc.) will have to rewrite native versions of their plugs.

    Pro Photo/Video:
    Flash storage readers (whatever flavor)-->Light Peak-->Mac. Mac-->Light Peak-->DreamColor-esque display. Storage pooling over 10 gigabit ethernet protocol over Light Peak to a 3rd party array. Basically a variation of what Small Tree is advocating right now. Render to a 3rd party farm via Light Peak.

    Anybody Else:
    Apple truly does not care about your pro needs. This is the company that bought Shake and then killed it. In fact, he doesn't really care about pro users at all and never will. Steve is about bringing art to the masses, rather than bringing the machines that make art to the artists. And most importantly, he will never build the gaming rig of your dreams. Or care about cutting edge graphics of any kind.

    So, if the above is a problem for you, I recommend you enjoy whatever speedbumps 2010 brings for the Mac Pro and xServe!
  2. PaulD-UK macrumors member

    Oct 23, 2009
    That sounds about right.
    Except: I don't think Apple will take control of what's outboard of the LP fibre, so there seems plenty of scope for third-party suppliers to give Pro Apps the connectivity and distributed processing that their users require.

    Not sure if SJ will want to run the possible risk of Michael Dell's derision by fitting out the cloud farm with 'Dell-type' non-Apple-branded boxes...
  3. maghemi macrumors 6502

    Aug 7, 2009
    Melbourne Australia
    Interesting Speculation.

    Time will tell if you're correct.
  4. Icaras macrumors 603


    Mar 18, 2008
    California, United States
    That was the best, most enlightening thing I've ever read from a "newbie" post, and I only meant that in context to this forum. Usually, newbie posts just pop up to ask when the next machine will be out. Its also nice to hear from someone who has observed the man for that many years as you have. Personally, I am relatively new to the Apple scene and its always interesting to me to see how the man goes about running the company the way he does.

    Anyways, I love the whole premise of Light Peak and where it may lead us in the future. It is way simpler to connect something via a cable than to open your computer up to plug in a PCI card. That I can say, I most certainly am excited for. With this technology, Steve Jobs will finally have his way with the iMac and the Macbook line being the all or nothing Apple computers. What you say makes a lot of sense to me, given Apple's current market trends.

    However, I don't like the idea of Apple turning their backs on professionals though, even if we have to admit that its a very possibly future. But if Apple can succeed in driving this industry to the ground with Lightpeak, then why would it matter and how would it compromise their professional market? I believe if Apple is successful with Lightpeak, then device makers will just adapt their hardware anyways. Users will just follow suit. And consumers, prosumers, and professionals can all be happy. Right?
  5. gugucom macrumors 68020


    May 21, 2009
    Munich, Germany
    I believe that the Mac Pro is doomed as a product. Whether it will be at Sandy bridge or the following tock is difficult to predict. Fundamentally it doesn't fit into a consumer company and can only earn the required margin if the price keeps being pushed beyond what the traditional customers would tolerate.
  6. cal6n macrumors 68000


    Jul 25, 2004
    Gloucester, UK
    1) No you don't.

    2) Everything based on your first assertion is pointless.

    That is all.


    Maybe my post was a little brusque, but you are flat-out wrong.

    Pundits and armchair technocrats have been predicting that Apple will axe their pro-level machines (and even that the tech industry will axe workstation-level machines en-masse) for at least 10 years now. It hasn't happened in that time and I can't see it happening in the near future. The tendency of the quality and quantity of the "work that can be achieved" (for want of a better term) in the creative industries to expand to consume the available resources means that there will continue to be an appreciable market for a pro-level Mac and the infrastructure to support a significant number of those pro-level Macs at the company level. I'll grant you that pro-level Mac is likely to use the latest technological advances, such as massively multicore processors and optical interconnects, and it may be significantly different from what we're used to, but it'll still be a Mac built for Pro-level use.

    Let's face it. If it's going to be good enough for Pixar (and it will), it'll most likely be good enough for the rest of us!
  7. AZREOSpecialist macrumors 68020


    Mar 15, 2009
    This is interesting speculation, but there will always be viable applications for the newest technologies in a more open, upgradable platform than the iMac. No matter what technology comes out suitable for iMac, there will always be a better, more expensive technology suitable for Mac Pro. I think we'll see Mac Pros for the foreseeable future.

    A more important question is whether Mac Pro is a profit center for Apple. If it is, as an executive Steve Jobs would be acting against the best interests of Apple to eliminate a profit center for no reason other than his own sense of aesthetics. Remember, Steve Jobs has a Board of Directors and shareholders to answer to. And in this economy what company would intentionally kill a profitable product line?

    If Mac Pro ceases to become profitable, that's when I'll start worrying.
  8. TheStrudel macrumors 65816


    Jan 5, 2008
    People have been posting this for a long time. If you're going to do another one of these "Mac Pro is dead soon!" posts, substantiate it with something more than "I think that...". You know, like actual evidence instead of more predictions of the sort Gruber loves to call "Claim chowder" after the fact.

    They'll stop making Mac Pros when they can't sell a million units of them any more. While it's true that they're selling far more notebooks than desktops, they haven't released sales numbers by individual categories. Mac sales are up overall.

    To pigeonhole  as "only a consumer company" is to make another sort of logical fallacy that there isn't real evidence to back up, just observations that they changed the company's name and sell lots of things that aren't computers.
  9. nanofrog macrumors G4

    May 6, 2008
    Very plausible, given the current direction Apple's taking.

    I don't think the workstation market is doomed, just the MP. Intel's direction of high core counts is aimed squarely at cluster/cloud computing. That's enterprise territory, and the independent/small entity creative houses won't be able to afford such systems. WAY too expensive.

    We'll see what's happened with Nehalem continue. A high-end desktop part that mirrors the features of the Xeons, just ECC support is missing. Which isn't a bad thing, as it's not really needed for a workstation (unless it's located within a nuclear facility, high energy physics lab, space vehicle, military,...). This would actually lower the cost a bit, as standard non-ECC memory is cheaper. We're about to have 6 cores on one chip, and 8 won't be too far behind. So a single CPU system will be able to take over that duty quite nicely. Combine it with non-ECC RAM, and it would even cost less in terms of $/performance.

    But given Apple's high profit margins and the rise in CPU costs (those high-end desktop chips won't exactly be peanuts), would push the MP out of the financial range or willingness to spend by the professional market.

    What I'm not sure of, is if they'll try to force cloud computing of workstation loads in the next few years. Bandwidth would be a problem right now, as the infrastructure just isn't there yet, and doesn't appear it will be in say the next 5 years.

    Even the large houses will look to switch systems if the price is too high. Though they have large budgets, they're not unlimited either. It's a business, and the accounting depts. would have a fit, as would stock holders as it would reduce profits.
  10. nanofrog macrumors G4

    May 6, 2008
    There's truth in the fact they have switched to a consumer oriented company though. Look at their product lines, and what keeps being added....

    There's more development in the consumer computers (i.e. introduction of the Air = new model). Then there's the more recent portable devices (iPhone, iPod,...), and the apparent introduction of a Tablet. It's all consumer systems/devices.

    And it makes sense. There's more money in it. Higher sales volumes, and high margins = more profit. The workstation market is tiny in comparison. Apple won't even give specific numbers for the MP or XServe.

    Sales have been up overall (profits), but it's from the consumer devices and systems. Just look at the recent marketshare of the iPhone. ;)
  11. goMac macrumors 603

    Apr 15, 2004
    Light Peak is still going to be way too slow for a lot of things traditionally on cards.

    For storage and other high latency situations, like input, it will be fine.

    For stuff like graphics and more demanding real time sound output, light peak will not work.

    Moving anything off the motherboard comes with a built in latency/speed hit. You not only have to go through several additional chips to make it off the motherboard, which takes time, but the signal has a longer physical distance to travel.

    Apple and Intel have been trying to get away from this problem as much as possible. Nehelem moves the memory controller onto the processor, Apple physically moved the RAM closer to the CPU on the new 09 Mac Pro, and with the new i5 mobile chips, Intel actually moved the graphics card onto the processor. Apple and Intel have been doing everything they can to move devices closer to the processor, not further away.

    With that in mind, it's obvious the Mac Pro isn't going anywhere, and card slots are here to stay. Light Peak, while it could work as a display wire, is totally unsuitable for graphics processing. Considering Steve was the one who ADDED PCI graphics cards originally with the B/W G3 as a stock option, he's not taking them away.

    But regardless, Light Peak is not suitable for replacing everything. It's high bandwidth, but also still high latency.
  12. Icaras macrumors 603


    Mar 18, 2008
    California, United States
    High latency? Do you have evidence to support this? I run my audio work off both firewire 400 and USB 2.0 devices, and while I'm well aware of latency issues, they are very much negligible. At least for audio, the vast amount of 3rd party audio devices out there use either Firewire or USB. Recording audio is certainly doable at this stage of digital audio technology. If it weren't, I doubt the entire USB/Firewire audio device market would have become this successful to begin with.

    I personally think we're already at the brink of overcoming any noticeable latency issues. And as far as cables go, Lightpeak should do everything that firewire and USB can do and then some.
  13. AZREOSpecialist macrumors 68020


    Mar 15, 2009
    Eventually we will hit the ultimate bottleneck - the speed of light. Once we are there, we'll know we can't go any faster until Apple switches to Intel's latest Dirty Whore architecture. It's rumored that Intel's new flagship will contain an on-die WHI (Worm Hole Interconnect) interface that will allow date to flow into and out of the CPU at 10x the speed of light through microscopic wormholes. This will be the biggest leap in computing since the PC Jr. We will still be asking Apple whether iTunes will cause these new CPUs to hit fusion temperatures, and Apple will continue to say it doesn't comment on unreleased products.
  14. VirtualRain macrumors 603


    Aug 1, 2008
    Vancouver, BC
    Hmm... I agree. In fact I was postulating a similar demise of PCIe here in another thread recently...

    However, I disagree that going solely with light peak is the path to minimalism... At least in the short-term. The problem with going light peak in a cold-turkey fashion is that you force everyone, including Steve himself, to have half a dozen out-board converter boxes so the new Light Peak enabled Mac is still useful and can connect to the outside world. Who want's a half-dozen different converter boxes?

    So while Light Peak might be the only interconnect on future macs, it won't be a cold-turkey switch.

    I also agree that the workstation market is disappearing. Many creative professionals can do their work now on a high-end laptop, whether it's video, audio, 3D, CAD, you name it. While any studio might have a few workstations around, it's more likely to equip users with high-end laptops and network accessible rack mounted HPC's for the real heavy lifting.
  15. nanofrog macrumors G4

    May 6, 2008
    I agree here, as it's only good for 1.25GB/s, and up to 12.5GB/s in a future revision.

    That's fine for an external port for HDD's, but for other uses, no way. It even requires PCIe lanes to connect to the system, as will NIC's at a minimum. Audio may go PCIe out of simplification, though that much bandwidth isn't needed for the basic audio devices used.

    It's aimed at graphics though, but will remain to be seen if it takes hold.

    Dual Link DVI is only 7.92Gb/s, so it may be fine, depending on the bridge chip designs.

    With Audio, it could be a mess. RME's HDSPe AIO for example, is only a 1x PCIe 1.1 lane card (250MB/s), but the latency could be problematic if it's not addressed properly in the bridge chip. If not, mulitple signals could have timing problems on top of the software aspect (variable layers).

    PCIe's not going anywhere just yet. It's still needed to connect devices on the board to the system. Not just Light Peak, but 10G Ethernet NIC's as well. Other devices may go PCIe at some point for simplification purposes.

    And those with existing systems that want Light Peak, will get a PCIe card to do so.

    Shrinking, yes. But given the direction of Intel stuffing as many cores as possible in the enterprise parts, the future of the workstation will be taken over by the high-end desktops (not in AIO fashion though), as the physical space will still be needed for things like RAID and multiple graphics cards for GPGPU processing.

    GPGPU will make a big impact for performance at some point, particularly for simulation use IMO, as it's far faster at FPU calculations. That means the boards will need more slots, as Intel's idea of placing GPU functions on the die has fallen behind do to technical difficulties. So the current high end PC boards are suited for such use, and they're already in existence. Some are aimed squarely at gamers, but a few are also offered up as Workstation boards (i.e. the Asus P6T6 WS Revo or Supercomputer boards).
  16. Dreamail macrumors 6502

    Jun 17, 2003
    Really good post!
    For Steve it was always about the democratization of computers.

    I would even go one step further and say that in the medium term Intel CPUs will be (mostly) phased out.

    With the acquisition of PA Semi and the investment in Imagination Technologies Apple clearly shows an interest in creating their own CPUs.
    Sure currently this is for iPhones, iPods and iSlates. But the writing is on the wall. With more and more cores, full GPU features - while keeping power requirements really low, in a few years an ARM Cortex A10 (or higher) with perhaps 8 or more cores could easily offer 'enough power for the rest of us'.
    While providing 12 hours battery life.

    This would be feasible for all Apple laptops and iMacs.

    Imagine the iMac no longer just a 'desktop' but instead a huge 1/4" thick slate that you can use as a desktop in its cradle but also pick up and use on your lap or flat on a table surface, or just hang on the wall.
    And of course it would run an advanced 'slate' version of OSX, not the OSX we know today.

    The Mac Pro with 2+ high-end Intel CPUs might still be kept for developers or as powerful home server for people who don't want to use Apple's data center for online processing power (or run their own 'Mobile Me' version) - for as long as it is profitable at its ever increasing price point.

    And the Mac Pro will be the last Apple computer running Mac OSX as we currently know it on Macs.
    The transition from Intel CPUs will be a subtle but huge OS paradigm shift as well - where most computer users will have migrated away from the classic desktop OS - to the mobile/slate OSX using a gesture based, perhaps even 3D interface running applications partially locally, partially online without even knowing (or caring) which part executes where.
  17. Dr.Pants macrumors 65816


    Jan 8, 2009
    I am going to disagree with you for these points.

    While certain high-speed connectors such as Light Peak could eventually meet the bandwidth requirements of some functions externally, certain professionals will send work to other machines for processing. For instance, rendering a two-hour movie in 3k? I don't want to be sitting at my desk waiting for my workstation to render for days, and I'm sure you (or several other people) don't want to, either.

    If Apple continues to put out its professional applications and other companies like Autodesk make products for OSX, XServe isn't dead. The only reason XServe would die is if, for whatever reason, Apple decided to let other OSes do rendering (like Nuke does).

    Not to mention that if Light Peak goes SATA (actual bandwidth is less then theoretical by a significant number), RAID and GPGPU are always going to be throttled - moreso at least then if they were attached straight to the motherboard.

    You're right, Apple doesn't care for our pro needs, they care about their pro needs (still doesn't explain the Apple RAID card, though). I'm of the opinion that whatever work machines are used at Apple will be the ones marketed. Once Apple's design department switches over to high-end desktops, then Apple's workstation is SOL; however, in that case, one CPU only renders half as much as two, so the logical thing is to send work out to computing clusters.

    I'd be surprised if ARM processors could pull off what Intel and AMD are doing currently. I can't imagine Apple using them over ULV processors, just because of the extra overhead for actual engineering (since ARM requires some additional technical details and custom chipsets instead of using one already cooked up by Intel, or summat along those lines). Not to mention PA Semi would have to keep up with the continuous die-shrink and increasing processing capacity of other processors.
  18. Fast Shadow macrumors 6502a

    Feb 9, 2004
    Hollywood, CA
    It's kind of sad to see a company that once cherished the creative professional market (and would have gone under without them) completely rebuke and abandon them in favor of moronic gizmos.

    Apple peaked as a computer company in 2003 or 2004. It's been downhill since then, and will continue to be as long as Jobs runs the show. But Apple the gizmo gadget company will continue to thrive.
  19. nanofrog macrumors G4

    May 6, 2008
    Markets shift,... and profits will outweigh any loyalty in a heartbeat, assuming there was any to begin with.

    That's where the $$$ is for them. The workstation market is microscopic in comparision to consumer markets (computer and device).
  20. SamuraiArtGuy macrumors regular


    Jul 13, 2010
    Eastern Panhandle, WV, USA
    Wow. Rather distills some of the thoughts bumping about in my head for a while.

    I have felt that the Professional Creative has been neglected by Apple for a while now, and that the Mac Pro is not a particularly high priority for them at the moment. With Apple's current overwhelming push into the consumer computing market, their interest in professional and especially professional creative users has waned as we're not nearly so important a market segment to Apple any more.

    Note for the purposes of this comment, you can pretty much read "Apple" and "Steve" interchangeably.

    Few CEO's have been more dominant in imposing their vision and preferences over a product line. And yes, Steve IS absolutely a minimalist, and this drives all their design decisions, for good or ill. And of course, pro users, and creatives in particular, often have need for features and ports that are directly opposed to some of Steve's aesthetics. Ports keep vanishing off the MacBook Pros, and pro users clamor for them back. Pro users have for years cried for multi-button mice. Several early "Hockey Puck" mice went DIRECTLY into the bottom desk drawer, replaced by cheap Logitech devices that did what I needed with PhotoShop and Quark(at the time). Steve keeps trying to chop the numeric keypad and full size function and arrow keys off the default Pro keyboard... but some of us do like, and a few of us need, them. The new player for Quicktime X neatly tossed a stack of useful pro-features to "tidy up" the user interface. Video creatives cry for Blu-ray support. Steve doesn't like it. SOL. Install your own blu-ray, monkey boy. ... or get a Windows Machine.

    The last few years in (my local) Apple Store, there is precious little Pro Gear in the shops. But there is an entire WALL of iPod accessories. Another WALL of iPhone cases. Genius Bar techs know a lot about iPods, but can't ferret out a bad Ram Module on a G5 any more. And I haven't seen a wide format printer in an Apple Store in YEARS. And September to the holidays, it's just not worth going into the places, unless you want to shop iPod/iPhone Central, with one or two Mac Pros sitting neglected in the corner of the store.

    Last fall, Apple was pushing the 27" iMac to Design and Video Pros as a viable machine. And if your budget is tight, it's compelling. You can get the 27" iMac, Nicely tricked out, AND a MacBook Pro for the cost of the bottom Mac Pro and Display. Or you can use the savings for the Adobe Creative Suite upgrade... (yes, a snark ) ... or an iPad... or iPhone.

    I think this is certainly a possiblity. Apple is definitely pushing in that direction.

    While the faithful are holding their breath every Tuesday, I'm keeping my expectations in check. I am not seeing a Mac Pro update until at the very least after the annual Septembers round of iPod rollouts and the G2 iPad, and I'm reasonably certain that the iMac will be refreshed before the Pro. Part of it is of course the market they're currently pursuing, computing consumers[/] as opposed to Creatives The iPad, and iPhone are superb media consumption devices. But I am not doing any kind of typesetting, print quality photo retouching and prepress work on an iPad, despite some very clever third party apps out there.

    Back in the day, "democratizing" computers meant putting advanced computing technology in the hands of Creative people, not just programmers and engineers. But now the tide has shifted, and "democratizing" computing is putting the tech in the hands of the widest (affluent) consumer market possible.

    The Mac Pro used to represent the pinnacle of Apple's technology, the most powerful and muscular iron they can build. So there was always a bit of bragging rights involved. But now the iPhone and iPad represents Apple's most advanced tech, the focus shifts.

    With Intel's next chip generation rollout looking a bit murky, despite tantalizing peeks at Sandy Bridge, what might be next for the Mac Pro is just as clouded. I don't think that the Mac Pro and XServe is up for imminent discontinuing, but they're certainly don't seem to be at all high up in Apple's to-do list. They are certainly not promoting the machines in any significant way.

    But you never know with Apple. They do still have the ability to surprise us. As for me, will try to nurse a few more trips around the block with my faithful and sturdy G5 tower, till I start getting Creative Suite files from clients or vendors that I can't open.

    That will be the $600 upgrade that costs $3500...
  21. Ravich macrumors 6502a

    Oct 20, 2009
    Portland, OR
    Ending the high end desktop mac line means pushing pros away from other apple products. Why would Apple do this? Why would Apple want a create pro industries where there is no room for apple products, given technology's growing presence and utility in the professional world?
  22. Hellhammer Moderator


    Staff Member

    Dec 10, 2008
    Why is there octo Mac Pro if quad core is all what Mac user needs? Gates said 384KB of RAM is all what a home user will ever need...

    Apple has fairly good pro apps and there are plenty of people who need the speed of Mac Pro for editing purposes. Bitrates are increasing all the time and better hardware is needed.

    Yes, I did not read the OP or any posts in this thread so ignore me if I totally missed the point but the topic should tell me all I need
  23. Roman23 macrumors 6502

    Jun 12, 2010
    Where is your proof that this may end up the fate of the mac pro?? How do you know?

    You honestly really believe that this maybe the end of the mac pro??? Where is your substantiated proof of this?

  24. TheStrudel macrumors 65816


    Jan 5, 2008
    I get the desire to panic, I really do.

    But everything we've seen thus far only argues that it's too soon to call.

    Abandoning Mac Pros is walking away from a lot more money to be made from Final Cut Studio, because that audience will buy one. To eliminate the rather modest workload this machine entails and profit gained from it wouldn't make a lot of sense.

    Less design effort goes into making the next Mac Pro than the next iMac, unless they actually do have a huge revision coming.

    However, given that the Mac Pro is marketed as a powerhouse machine,  may just be holding off on updating until they feel like they have something more substantial to put in it.

    Given how underwhelming the last Final Cut update was, I actually agree with that sentiment.

    Oh, and I know this may break your heads to read it, but there is actually more to Apple than Steve. And that's been true for a long time. He actually listens to some of his engineers, if you can believe it. If you think he designs or makes every single decision about every product they make, I've got a few bridges to sell you. Wouldn't surprise me if he ignored Mac Pro development completely until approving the final release or telling them to tweak it a bit.

    I'd rather they waited until there was something much better to deliver. And I'll bet dollars to doughnuts that everybody on this forum, even the most vociferous whiners, would also rather wait for a substantial update than get an incremental, unimpressive one.
  25. Roman23 macrumors 6502

    Jun 12, 2010

    I agree with you 300 percent!


Share This Page