Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
New Mac Pro 2014 release?
- Thread starter McNismo
- Start date
- Sort by reaction score
Nobody outside of Apple knows for sure. This might help you make a decision:
https://buyersguide.macrumors.com//#Mac_Pro
They won't release one this year, it's simply too late now. They will wait until Q1 or Q2 next year. I guess it depends on how long you can afford to wait and it's not like the new one is going to be 400% quicker, expect ~10%. Weigh up the small increase in performance (that you may or may not see depending upon your workload) to having the machine now.
Do you think there isn't a big speedbump?
When I read the following article, I doubt to buy it now.
https://www.macrumors.com/2014/08/14/intel-new-xeons-mac-pro-upgrade/
When I read the following article, I doubt to buy it now.
https://www.macrumors.com/2014/08/14/intel-new-xeons-mac-pro-upgrade/
Last edited:
Hadn't seen this page before, some really interesting insights there.
As for when will they update, I wouldn't think too soon. Perhaps a 5K display first? Or launch the display and Mac Pro at the same time. Hopefully they will change the design so that it is internally expandable again, massive shot in the foot for the 6, 1 for me. Also unfortunately I can't see there being any Nvidia CUDA option again which is a no go for a lot of professionals.
The Mac Pro could be refreshed anytime if the hardware is there. It's a different market than the rest of Apple's offerings.
The hardware is not there. It's not any different from anything else as ultimately everyone is dependant on Intel and they are delayed.
More likely it is software/firmware that is not there. Both Intel (Xeon E5 v3 ) and AMD (GCN 1.1 : Tonga/etc) have updated hardware that would work in basically the same Mac Pro core design with a few tweaks. Apple tends to be behind the curve on GPU drivers availability ( Nvidia/AMD tend to work on there larger markets first and roll out OS X variants later. ). The rest of the FirePro x100 cards besides the 8100/9100 cards aren't finished driver wise either. The Apple derivative variants are likely going to trail a bit.
More broad support for single stream 4K monitors wouldn't hurt new Mac Pro pro models either. A wider set of those monitors are probably coming around January (CES 2015 show).
Similarly, Apple has the broad spectrum 10.10 (Yosemite) release to worry about right now. Apple certainly doesn't need an iOS 8.0.1 debacle happening on folks production Mac Pros. It would not hurt to have a Mac Pro release that didn't need a firmware/OS update on day zero of availability.
Waiting for E5 v3 to go into high available, volume production and similarly for the GCN 1.1 also reach that state would help. Apple really doesn't need another "almost 2 month backlog" launch. A couple of weeks and the beginning is kind of an Apple thing, but extended scarcity tends to piss-off folks as much as fuel the buzz.
Apple shouldn't need until Spring to get a revised Mac Pro out the door, but they also don't urgently need to release a new one over next 2 months either.
The current one wasn't commonly available until April-May this year. That wasn't that long ago. They will probably want some pent-up demand to build up. Also likley to invoke of bunch of blacklash against a new version ( several threads already popping up... "Haswell not needed.. Apple should skip (so that my brand new MP isn't sunset so fast)" )
Last edited:
Pragmatically this current Mac Pro 2013 model was really a Mac Pro 2014. It wasn't widely available until 2014.
A 12 month offset from the actual/pragmatic availability date would make for an "Early 2015" Mac Pro; not a "Late 2014" one.
Apple had dragged its feet for years after the 2010 MP (with the 2012 model essentially being a "it isn't dead' statement; not a we have been diligently working on something one). So getting back to an approximately 12 month cycle would be a dramatic improvement for Apple. Getting to a less than 12 month cycle is behavior they haven't engaged in for very long time (more than several years. ) .
I would not see a new Mac Pro Refresh until Early 2016 or at least when the 5K Display comes out.
Apple right now is focus on Consumer Products (iPhone,iPads,etc..), there are not to focus on the Pro Market because the Pro market does not move units.
Also when you buy a Mac Pro, you might keep it for 5-8 Years, Apple has to wait a while for new money, Its not like the iPhone/iPad which moves a lot of units/ has high markups and consumers usually upgrade every 1-3 Years.
5K is not a panacea. 4K monitors work pretty well. The newer single stream ones will probably work better over a broader set of offerings for the Mac Pro.
There is no good reason to hang up a Mac Pro release simply based upon the external monitor. There is little to no indication right now that Apple is even going to do anything to get back into the monitor business. Effectively Apple is in the docking station (with embedded monitor) game.
The Mac Pro 2013 released with Apple acclaimed 4K capability and no Apple 4K monitor. It would not be any shift in strategy at all when it comes time to dog and pony show 5K monitor ability to rely on 3rd parties.
In early 2014, Mac Pro demand outstripped Mac Pro supply/production. The Macs don't have to sell in the same unit numbers as iPhone/iPad. That is just misdirection. What is far more important is that people are buying them in sufficient (and preferably growing ) numbers. Overall Mac sales units are up, which very likely means Apple won't be dropping or backing off Mac R&D any time soon. Mac Pro's numbers are up but there was such a long drought prior that Apple will probably want to probe to see if that is longer term trend.
iPod sales are down and iPods are getting little to no R&D. Apple canceled the iPod Classic in part because "nobody" (so few) were buying it. That isn't particularly so few relative to iPhone... that was so few compared to how many they used to sell. ( folks may have swapped purchases of iPod for iPhone but again that is just following the market changes ).
It isn't so much that it is 5-8 years as much as the demographic change from folks moving from 3-6 upgrade cycles to 5-8 ones. Folks who don't need performance are an odd mismatch to the Mac Pro.
As large a set of folks coming out of a 5-8 cycles as 3-6 years ones is the same arrival rate of new purchases. Filling the pipelines is already done. It is the shifting of pipeline lengths over time that will stall demand for new boxes.
Shrinking demand is what will heavily influence Apples focus over time.
Last edited:
The dual CPU market isn't growing substantially. Increasing numbers of cores being packed into single packages negate the need/demand for multiple CPU packages. What folks want is cores, not packaging. A 12, 14, 16 core progression is going to cover an increasingly larger spectrum of users.
Neither is the "bigger box under/beside desk" market a growth fountain at the moment.
The second internal SSD is more a mismatch at the internal components level. The shift entirely from SATA just isn't reflected in the Intel support chipset variants yet. Even the new generation that comes with E5 v3 (and v4 since they probably will be sharing same chipset). [ The C600 and C610 chipsets that support the E5 v1-2 and v3-4 respectively are not a single product offering. There are variants that vary along SATA/SAS support. Longer term that is likely going to expand to include SATA/SAS/PCI-e support when it comes to storage. ]
Apple could have put a SATA based SSD in the Mac Pro but then they'd asymmetrical. I'm sure there would be moaning and groaning about that too.
PCI-e SSD blades , Thunderbolt , and 10GbE isn't particularly well provisioned to do all of that with Intel's current chipsets. When they do provision for that kind of concurrent PCI-e bandwidth loading then will have higher likelihood that will see Mac Pro's with dual internal PCI-e SSDs.
My problem with that is no matter how great a single processor can get, you will still only be limited to one as other systems will be capable of twice that. I'm fine with my 2012 MP right now, but in a couple years I think Hackintosh will be my only option. Power users demand options and upgradeability. Thunderbolt does not solve all my needs. I would gladly trade Thunderbolt in exchange for my PCI slots.
That is your problem. Your problems aren't Apple needs. Nor are they particularly illustrative of a substantially large and growing market. There are folks who need two package solutions. The large, dominate vendors in the workstation space (who are not Apple; HP, Lenovo, Dell ) will probably fill those needs over the short and intermediate term.
Again there is a difference between all of your specific needs and the overall market needs. The sweeping generalization from yourself into all power users is smoke.
As long as the Hackintosh market is some 1-2 percentage (or lower) of deployed/active OS X instances, it just reinforces why Apple should just bypass those subsegments.
OS X pretty much ignores NUMA issues. Large core count single package CPUs will take on small scale NUMA issues that likely are in same ballpark that the older Mac Pros tip-toed around. Explode the core counts with multiple CPU packages and sweeping those problems under the rug likely is going to turn into a noticeable bulge over the long term. It is natural counter balance longer term Hackintosh growth in that subsegment.
The Hackintosh market isn't going to explode into some super high growth subsegment any time soon.
Are there any significant updates to the chipset in the Haskell-EP based Xeon's though? Rather than a small speed bump Apple would probably wait until they can introduce Thunderbolt 3 and more PCI lanes.
In line with this, the Xeon E5-2699 V3 has up to 18 cores:
http://ark.intel.com/products/81061/Intel-Xeon-Processor-E5-2699-v3-45M-Cache-2_30-GHz
It is fabricated using 22nm lithography. Intel is already using 14nm for some products. In theory putting that Xeon on 14nm would increase the transistor budget to enable 36 cores, with plenty left over. With that potentially on the horizon, this further diminishes the need for multi-socket workstations except at the extreme high end.