Tim Cook confirms updated Mac Pro coming in 2013 - Page 2

What do you mean it is pretty obvious that the petition was a waste of time.    Unless you call this "new" Mac Pro the type of machine Mac Pro users where expecting.

It might be a waste of time answering the questions below.    I will give it a shot though.

 

First; the current Pros are crap.   Basically you have I/O that would have been great 3 or 4 years ago but is now an impediment to fast operation.     Second; no one cares about Final Cut.    Third; if you don't know why paying a premium price for a 3 year old GPU card is a problem you need to get a life and a bit of an education.

 

The problem is it is a workstation costing thousands of dollars as such people expect modern hardware for that cash out lay.    Think about it a bit why in hell would you spend multiple thousands of dollars for a machine that doesn't support modern I/O?   We can focus on USB 3 or any other part of the system if you like but really why would you lay out cash for hardware that has been dated for three years now?   Especially a machine that is an investment and has to last.

 

AS to Pros loosing money it should be pretty obvious, if the competition does the same work in one quarter of the time you will suffer as a Mac Pro user.   It is simple business where time is money.   Your mi 2010 line up is technically very dated.   In many cases that line up will be out paced by todays laptops and certainly by run of the mill desktop machines.   

 

Really I don't know what you are harping about.    There is no defense for the current Mac Pro that I can think of.   Why you would choose to emend this move by Apple is beyond me.   Consider that this machine doesn't even meet the desires of the most conservative of members on these forums.   Last week few would have believed that Apple would have seriously tried to deliver such a machine.

The timeframe of late 2013 makes me think this will be the big overhaul. While Ivy Bridge Xeons would be due at that point, I really don't think it would be wise of Intel to bother with Ivy Bridge Xeons. They have 22nm already so they might as well jump right to the Haswell architecture.

10-core/20-thread single CPU, 8" Cube design, 512GB SSD blade, 3x platter bays, fast GPU only upgradable from Apple using HSA architecture, 6x Thunderbolt ports 20Gbps each, 4x RAM slots (up to 64GB RAM), Thunderbolt daisy-chaining for compute sharing CPUs and GPUs.

No way they'd wait 3 years and build another giant box.

It is something the idiots in petition villa don't grasp. Tim basically said this is the end of the road for the Mac Pro and that they have something planned to replace it. That is good of course but it still doesn't justify the current Mac Pro update.

 

I am not sure why anyone cares at this point. The Mac Pro is two years old and Apple is asking their customers to wait another year? 

 

If you are running a business that uses Mac Pros and the professional software that goes along with it, it is time to start thinking about Windows. For all we know, Apple's new Mac Pro will be the hardware equivalent of FCP X. If history is any indication, we will still get a limited selection of GPUs and they will likely be out of date. We won't have a choice either, it will be whatever manufacturer Apple decides to go with, ATI or Nvidia.

 

And after the next one comes out, how long will the next hardware update take? 12 months? 18? 24? Why any professionals would invest in such a platform is beyond me.

 

-kpluck

What was that about nine camera views running at once?

I don't think Retina 15" has any more "screen real estate" than the regular 15" MacBook Pro. Just sharper. Like the iPhone and iPad. The 17" still has more on the screen.

 

(this is a reply to a post on page 1 about the Retina 15")

I'll venture a guess and say MAJOR updates at NAB next year.

Mac Pro with a new form factor, a headless iMac (just 'Mac'?) and new iMacs.

Which implies that the current Mac Pro is a dead platform.

 

I'm in agreement that a major overhaul of the machine and even the concept of a high performance computer is in order.   They almost have to innovate massively to pull themselves out of the mess they just created.

 

While we could speculate for another years as to what this box will look like, it is the current box that is the problem.    I just see a rapid decline in sales.   Further the complete and utter disregard for the desktop shown at WWDC doesn't look to good at all from the standpoint of a customer.   AT best we will be waiting for another 6 months and likely far more for even a hint as to what Apple has planned.   Looking at it as a span form the last Mac Pro update that will be well over 4 years of time for technology to pass Pro users by.   That is the big insult.

 

As to the machine you outlined below, I was actually hoping to see such a box replace the Mac Pro this year.   The technology is there to do so.   Granted Intel has some interesting technologies up their sleeves that could make for a better box in 2013-2014.   For example if they get their new 3D memory modules on line.   In any event we could easily be talking another 12 month wait here and that would mean that very stable hardware will be in the hands of the competition for that period of time.

 

I really think Apple blew it big time with this screw up.   At the very least they should have done something with the other desktop machines to demonstrate that they have a clue with respect to desktop user needs.

Your points are very valid.   However the issue with the GPU is more complicated I'm actually expecting the next machines to have the GPU integrated right on the motherboard.   It is the only way to move technology and performance foreword.   We can only hope that the GU isn't to dated.

 

By the way form the software development standpoint having a fixed GPU for a specific model is a huge advantage.   It means time spent on optimization is less likely to got to waste.

Just add a large Apple external screen for when you are a desk and you're golden. That's what i did.

Repeated from another forum I posted this in:

 

Why in the world does Apple act like they can't "afford" to update more then what is making the most money at the time? They say they believe in making the best products possible but what they really mean is making the best products possible with the highest profit margins. How in the world is updating the iMac or Mac Pro detrimental to them even if it is only for their "hard core" mac users? These "hard core" Mac users are the ones that kept them alive before the iPhone and iPod.

Yeah, the [arguably] best multicam implementation out there running full bore on a laptop... And displaying a full 1080p video running in the upper right corner of the screen
Edited by Dick Applebaum - 6/12/12 at 5:51pm

This pro offering may just be a necessary peace offering to the pros... Obviously, Tim sought out a dissatisfied user to respond to, in order to assure him that Apple is not going to abandon this very important market.

Even if it costs Apple $1 million -- it is money well spent... Advertising!

Showing the flag!

I'm not sure why (most) everyone here is assuming that the "something really great" Cook is hinting at will look anything like an existing Mac Pro.   If all they wanted to do was to build a newer Mac Pro, there's no reason for that to take a year and a half to put together.  Meanwhile, the market for Mac Pros is clearly declining (I have one and love it, by the way -- no hating here!) and, if I were Cook, I'd be looking for a way to provide lots of computing cycles to my high-end customers, but in a way that might also pay off for other customers in other ways.

 

Perhaps give a Mac Mini a bigger processor, an SSD drive, and a Thunderbolt interface, and then let people hook together as many as they need?  Two for my home server, ten for my video rendering box at work.  Maybe open up Apple's cloud service so that I can offload compute-intensive jobs in a way that seems transparent/invisible?  There are lots of ways to give people lots of computing cycles other than to stick a bunch of stuff in a big box under their desk.

 

"Think different", remember? :)

The bottom line is this comes down the number of units sold, the costs of parts and production, and what Intel wants to do.

 

For a MacPro, Apple needs as many cores as it can throw at multi-threaded pro media apps.  To get more cores you have to go multi-processors, or Xeons.  A single i7 might actually be as fast or faster than a Xeon at many tasks, but when you throw in dual processors the Xeons win when doing heavy processing.  The Xeon processors Apple use in the MacPro are older version, but even the latest Xeons are not Ivy Bridge, they are Sandy Bridge.  And the increase in performance of the newer Sandy Bridge Xeons is not a huge gain.  Intel has been focused on die shrinking the Ivy Bridge i-series chips because they can sell millions more of those than server/workstation processors.  

 

The next part of the equation is the chipsets and sockets needed to support the CPU's and provide I/O.   Again, Intel holds the cards here.  I believe the older Xeons the MacPro has use the LGA 1366 CPU socket, and the newer Sandy-Bridge Xeons use a newer LGA 2011.   So to get a marginal increase in performance, along with some power and heat reduction, Apple would have had to design a new motherboard around the LGA 2011.

 

For massive volume products, Apple has no problem creating their own chips and chipsets and having them fabricated.  For a product like the Mac Pro they are going to rely on the 3rd parties as much as possible for standard designs.  Right now, it appears that none of the chipsets out there that support multi-processor Xeons provide ThunderBolt support.  So Apple would have to add support for Thunderbolt.  

 

Next year a new die-shrink Ivy Bridge Xeon is expected, and the support for Thunderbolt will be more wide-spread.

 

So bottom line, Apple looked at its options and said "we don't want to invest a lot of money to create a new, marginally faster computer model that will probably need to be replaced need to be replaced again next year.  

 

So this is what they get for going with Intel...  delays on releasing new MacBooks because Intel was late with Ivy Bridge CPUs, and slow development on high-end Xeon lines.  But the alternative of going with PowerPC would probably have been worse when looked at over the long-term.

 

I would look for the NEW MacPro's next year when the Ivy Bridge Xeons come out.  Sadly, Apple could have chosen to at least offer faster graphics cards, but perhaps they figure that most pros will choose their own video upgrade options.  However, by showing weak support to the MacPro, Apple does not send a good sign to 3rd party video card makers to stay in the Mac market.

 

And now you know why Apple likes making its own chips for iOS devices.  

"we're working on something really great for later next year"

 

Translation - A man Mini with three thunderbolt ports, a 1TB SSD, and a high end video card.  From the Apple perspective, this is exactly the product they think is great for the high end market.

 

I don't see anything in Tim's email to suggest that the "Great Product" will be anything like the current Mac Pro.

 

 

 

GOOD. You really want the same thing regurgitated every year?

I agree!

How about modular boxes daisy chained together with thunderbolt and/or fiber optics...

Separate boxes with: RAM/CPUs; SSDs; GPUs; HDD RAIDS... mix or match these as needed to address current needs.

Apple already has software to manage this distributed computing system

My feelings exactly, but far better put.
Edited by anantksundaram - 6/12/12 at 6:29pm

Its not for me what my current mac pro can't do for me as how fast it can do it for me.  I do a lot of video editing on my machine and a lot of redering out to h.264, this is where the mac pro's shine, multicore support and hard crunching of video data is just what these machines are designed for.  The xeon processors are a workhorse processor, I can max out all my cpu's encoding video  and run them that way for weeks if necessary and have complete stability.  iMacs just are not designed for this kind of use.  Also the expandability of the Mac Pro is another area where professionals have to have a machine that can support multiple pci cards and expandability.  

 

Having said the above  pros need faster and faster machines to do there work and keep up, The current mac pro like has been said is two years without a refresh, and starting to lag behind.   Pros are people that make the things most users enjoy so apple needs to really keep them happy.  There is no excuse for not taking some of those billions of dollars apple has in cash and investing it in keeping pro developers happy with a current and speedy workhorse machine.

The new retina macbook pro has higher resolution but less screen real estate than 17'' MBP. Really, it has the same screen real estate as old 15'' MBP (2880x1800 pixels is exactly 4 times 1440x900) so 4 times as many pixels are used to render same size UI elements as on the old 15'' MBP. You are not going to see more of text in a web page or text editor on the new screen.

 

"The PC wars are over. Microsoft won." ...Steve Jobs

 

Just like any successful companyApple looks to the future and the future does not include desktop PCs. Scream or wring your hands all you want but Apple's future lies with mobile devices. Apple does not consider itself to be a traditional computer maker anymore. AT&T now considers itself to be a wireless company, not the old wireline company it has been for over 100 years. As has been said many times lately we are entering the post-PC world. Get used to the new Apple as that's where Jobs was headed when he retired to the spirit world.

 

The future doesn't include laptops. Desktops are going to see a resurgence. Laptops in the short term, but they'll be useless when direct-contact multitouch comes to desktop OS'.

 

I wouldn't be surprised if it's a vastly different beast to the current Mac Pro. I bolded your mention of Thunderbolt because it's when Thunderbolt gets off copper and onto optics that the fun starts and Apple can really make some big changes to it's workstation-class computers. "Pro users" don't mind change, but they do need support for high-end I/O.

Companies and individual pros have budgets you know. We buy new computers every once in a while, when it is justified or when budget allows. We will not put our money into a dieing platform or 2 year old computer that still costs exactly the same as the first day it came out. It is simply a senseless purchasing decision to do that.

 

This is not about current machine not meeting needs as much as it is about future planning and budgeting and knowing your platform of choice is going to be supported when you do need a new machine. Otherwise, we need to plan and budget for total platform switch, which is costly and disruptive. This is why typical Apple silence on their long term plans is so frustrating.

 

You don't get it, we get that. "Run Final Cut"? It's not about 'running' software, it's about how the computer is able to handle the codecs. Our '08 Mac Pros 'run' FCS fine, but are struggling big-time with current codecs - something that has really hit hard in the last 12 months. We have already upgraded to custom GPUs for our suites running Davinci Resolve for grading, and we'll grab one of the "new" Mac Pros just released as a stop-gap measure only.

 

Given that the top end Mac Pro is an expensive beast, production houses like ours not only keep these machines for as long as possible but also look to new purchases to be able to have a long life. We need them to not only come with more grunt to be able to support modern codecs (HD/2K/4K/5K) but to have up-to-date I/O, ie Thunderbolt and USB3 in order to work with and to be able to get massive amounts of media in and out as quickly as possible.

 

Is that how you see it? If so then I think you mistake the intention of the petition as well as the concerns of those using Mac Pros in a production environment. As much as getting a new model it was about getting some kind of indication as to whether or not they were going to have to jump from a platform that they have invested heavily into. The fact that Tim Cook responded means that it wasn't a waste of time at all.

Are you using FCP or FCP X?

 

What's the issue with modern I/O?  

 

1) Many of the Pro users I know are running FC or SAS HBAs and not USB3.

2) You can always add a USB3 card now.  Just like you have to add a FC card.  Mostly useful for 3TB external drives for sneaker netting...at least around here.

3) TB is great but most of the products are geared toward mobile and not workstation.  Many high end workflow TB solutions involve buying a PCIe chassis from Sonnet and sticking in PCIe cards...which you can just do in the Mac Pro.  TB really helps iMac and MBP/MBA users a heck of a lot more than Mac Pro users.  In fact, TB helps you transition from needing a Mac Pro for expansion to a smaller footprint.

4) The Mac Pro has not been dated for 3 years.  This is either a typo or a truly idiotic statement.

 

The old GPU does suck...hopefully a new one will appear soon.  If you need a new Mac Pro today it kinda sucks but you get to depreciate it 50% this year and you need it to get work done.  Generally the billable hours outweigh the cost.  Most Pro users will just wait till next year and hope it's actually EARLY next year and not late.

This new forum system can't quote properly at all.

What i was going to respond to is what the Mac Pro replacement will look like. I think you are a bit negative here in describing it as a Mac Mini as I'm really hoping for more. Note I said hoping, I've lost confidence that Apple has really good plans for the desktop.

I do expect a dramatic change in size for the pro replacement. If they put half as much effort into it that they put into the retina Laptop it could be one hell of a machine. That is if they keep an eye on the ball and truly design a machine for professional usage. There is a real and justified fear that they don't grasp professional needs anymore. Of course most professionals can't see beyond their own needs so this is no surprise. What they do need to do is to make sure they end up with a machine that appeals to a wide array of users.

 

What are you using today?  Harsh language?  Tell me you're using FW800 and I'll ask you to pull the other one.  What do you expect TB to do that 8 Gbps FC cannot?  Do you really think you're better off connecting to a Promise Pegasus via TB over a Promise Ex30 with 8 Gbps FC ports?   You'd really prefer to use an Aja IO over an Aja Kona?  

 

I guess everyone runs out of slots sometime but still I don't get it...yeah TB ports on the Pro would be nice but the PCIe lanes come from somewhere.  There are a lot of use cases where you might want those 4 PCIe lanes not configured as 2 x 10Gb bi-directional.

 

As for USB3, I guess the Mac Pro has enough CPU power that you'll never notice but it appears that many times higher end legacy interfaces run faster with less of a hit on the CPU.  Like say eSATA.

It is no different than in the hard econmy goods. Take a machine shop for example that has a CNC lathe capable of making one part at X speed and a competitor across the street with a different lathe that can make the part 4 times faster. Given equal hardware costs who do you think will make money on producing these parts.

What many people can't fathom here is that it is exactly the same deal with the Mac Pro. The Pro user might not be fabricating valve parts on a lathe but rather the parts might be video files, a structural analysis or a sequencing code. Whatever they are doing the complexity demands powerful computers and staying up to date. If the competition can buy 4x computer performance for the same number of dollars then he has a huge advantage over someone running an aging Mac Pro. Right now 4X is a realistic number though the cost might be higher.

In the end it is the old yarn about time being money. Time here being computer time.

Another way to look at this is the concept of a job submission. Let's say user X has to submit his jobs at the end of the work day to view the results the next morning. That is a job might take ten or more hours. Now user Y has a much faster computer that can run the same job in a couple of hours. So he submits the job before lunch, takes some time to do some banking or surprises the wife for some afternoon delight. User Y comes back to his work station after that long lunch and has several hours to review the job that just finished up. The review finds the need to tweak a few things so the job is resubmitted again that day. User Y is already well ahead of User X, a two hour turn around is giving him a huge advantage over user X. Realistically he can turn the job around three times in one day if he has too. The same job might cause user X three days of effort.

Now obviously the above is slightly contrived. However modern hardware and OpenCL type acceleration has caused such jumps in performance such that overnight jobs become a couple of hours of processing time. In the context of the Mac Pro it would not be impossible to find a workstation today that can run 4X faster on some jobs. It is a real concern if you are in business and have agressive competition.

 

Doh...erased this part in the post.

 

IMHO going the route with GPU-Xpander is better than the bump from a Sandy Bridge Xeon update.  That burns a full x8 or x16 slot but you can cram what?  4 Teslas in the desktop model?  With their own power supply?  Dunno, we don't have one as far as I know but I don't work with any of those guys any more either.

 

I'd also much rather have 8 lanes available for this over having them dedicated to 2 x TB ports on the Mac Pro.

In this case I meant literally all I/O. I mention USB 3 because of it's fast grow or acceptance in the marketplace. However if you ignore that at the moment you will see that all I/O on these machines be it memory, disk interfaces or PCI Epress lanes is ancient in computer terms. It is hardly smart to invest in something like the current Mac Pro when everything you buy for it is effectively old and slow. Sure many are but as many are running plain old SATA devices. Devices that could perform very well given something better than the current interfaces. Sure if you want to use a slot, I might suggest though that there are better uses for those slots. Especially considering it is about time to build such interfaces into the machine. Really I'm under the opposite impression. Well it would considering that the Mac Pro doesn't have TB ports in the first place. Realistically TB doesn't even do that. TB is not some black magic that conjures up unlimited bandwidth. It is hater a serial connection with everything that implies. Well do you call what we got yesterday an update? All we got was a pathetic bump of the processors. Not even the video card got touched which is slightly more than three years old I believe.
This I agree with, many will simply wait or defect. Nobody is rationally going to invest in this technology, especially when it is clear that the platform is dead.

 

Mmm...in a new buy scenario yes an updated Sandy Bridge Xeon is obviously a no brainer over the current Mac Pro.

 

In the scenario where I already have a 2010 Mac Pro you kinda have to show me that you're going to get 4X performance just by switching to Sandy Bridge.  Maybe I buy it because you're moving from PCIe 2.0 to PCIe 3.0...that's double the bandwidth to use on GPUs...but the scaling isn't generally linear and it's still only 2X and not 4X.

 

In a capex constrained environment am I better spending $$$ to replace the existing Mac Pro with the Sandy Bridge Mac Pro or fully upgrading the current one?   Is it better to add to your transcoding or render or compute farm than upgrade a dozen Mac Pros to Sandy Bridge?  Or hell, just buy time from Encoding.com?

 

That's not my line of work but I DO know it's a hell of a lot cheaper for me to buy time on Amazon EC2 for some compute tasks than the kit out my own server farm.  I can get a Cluster Compute Eight Extra Large Instance (60.5GB RAM, 2x Intel Xeon E5-2670) for a HPC job for $2.40/hour.  Which sounds so-so if you want to run one of these 24/7 for 365 days but some HPC use cases I'd like 1064 instances x 24 hours.  That would only cost me $61K assuming Amazon could actually sell me 1064 instances at any one time.  Not likely, but I have scenarios where I could use 100 compute instances for a monte carlo run.  Assuming the job takes a day that's still under $6K.

 

If I need to use CUDA I can buy a Cluster GPU instance with 2xXeon X5570 and 2xM2050 Fermis for $2.10 an hour.  Each card is still around a grand each.  I can get 1000 hours for the price of buying two of these cards.

 

I assume the economies of scale work about the same way with something like Encoding.com.

 

So no, I don't buy into Mac Pro shops are doomed vs competitors sporting new HP or Dell Sandy Bridge Xeon workstations.

This trickle of information really makes it difficult to recommend Apple's products to clients. They don't update the Mac Pro for two years and then give it a minor update. It's hardly a "Pro" product anymore. I gues they don't care about this stuff anymore: http://jeff-with-a-g.blogspot.com.au/2012/06/apples-high-order-bit.html

 

Where the hell do you get that?  Cook just promised a full update in 2013.  That doesn't preclude another spec bump before then. 

 

Mostly I expect a case mod in 2013 to allow for rack mounting.  Slapping Sandy Bridge into the existing MP is still a possibility.

 

A leak is an unsanctioned. This came directly from Tim. Doesn't get more official.

 

 

 

What about the Mac G4 Cube? Nope, still dead.

Mac Pro: still alive.

 

That's exactly what he's talking about. Executives talking about future models before they're out. This is a first for Apple. Ever. 

 

I don't like it at all.