People Who Complain About the New Mac Pro Misunderstand It

[barmann]

Not going to happen. If Apple loved making rectangles with slots they could just made one.

Look for Thunderbolt to succeed it has to get picked up by the overall PC market. Apple screwing around threatening to wipe out perhipheral vendor's investment isn't going to help that ecosystem grow. Apple has staked out a small subset of highly effective or highly necessary (due to possible premature port eviction by Apple. ) Thunderbolt devices.

Their display docking station is effective because it combines both things that TB does well video and PCI-e data transfer. Not just one. The TB dongles are necessary because a significant number of folks do use Ethernet jacks and firewire ports on their laptops even though Apple is nuking those. ( The iMac even drops FW which is kind of goofy. )

Worth quoting; much discussed already, yet important points still .

It really comes down to the present state of TB technology, I think .
Or rather the current availability, maturity and pricing of TB devices .
Love the 'display docking station' part , that's particularly weird .

TB dongles might disguise the impact of Apple's shift to TB; they are fine as a quick workaround for iMacs and Laptops, and there still is USB as a viable option for those .

With a MacPro, that's a very different matter .

I still don't get it how Apple think they can just throw this new MP into the existing workstation environment, and expect third party suppliers to sort out the rest .
And right away, too, this clientele waits for noone to get into gear .

 

[deconstruct60]

Apple's published product lifecycle support policy is for five (5) years and in the specific case of the Mac Pro, they know that its customer base is essentially atypical and will push hardware to its technology limits...which includes emerging technology becoming more affordable.

I loop back to 5 years in after clearing up "support"

Err, no. This has to do disconnect between what provide support services technically and the notion of "support"/"supported" meaning "being made to work somehow". Support services are consultative ( helping users use what they bought) , diagnostic ( helping users work through problems), fixing/patching broken things ( things that should have work because it was explicitly sold as having that functionality).

Adding new features is not support. Diving into 5+ year old firmware to add features is a feature addition. There are no new features being added to OS X 10.6 ( or 10.5 or any lower ones). If it was 1-2 years into the firmware's lifecycle then maybe if it was low risk. But at this point the regression testing infrastructure and span of OS X instances that are to be tested pretty much lined up for retirement, not bring new extensive run.


Technically Apple's policy is 5 years after discontinuation. This super long extended transition and shared motherboard over full tick-tock cycle pragmatically presents a big disconnect here. The 2010 (model 5,1) technically could start the clock on the 2009 ( model 4,1). Many of the single package Mac Pro even into 2012 variants are 2009 era CPUs.

Intel is actually going to pull the Xeon 3500/3600 from the retail market in September 2013. They are going to pull them from system vendors in 2014. Apple's 5 year window is not something they just pulled out of their butt. Intel, IBM , Oracle, HP lots of major players put down 5 year timespans so normal support. High end vendors have paid for support that rolls long but a 5 year window isn't really an "Apple only" thing.

...and yet because 16GB DIMMs can currently be used in a Mac Pro, it results in a Mac which can go beyond its advertised 64GB 'limit'...and pedantically can reach today that same 128GB that is being touted for the Tube.

To be blunt there s a bit of disingenuous spin here on both sides. The reality is that the dual package Mac Pro is dead. If want to an Apples-to-Apples comparison of new Mac Pro to old Mac Pro, then it is really single package variant versus single package variant.

It is not more any more right to compare the dual Package Mac Pro 2012 to a single package Mac Pro 2013 than it would have been for Apple to compare a single package Mac Pro 2009 to a dual package Mac Pro 2010 or vice versus ( single 2010 to dual 2009 ). All that does is more so highlight the different product SKUs as opposed to what has been actually upgraded.

So as sold and provisioned can the single package Mac Pro 2012 reach 128GB today? No.
Will the MP 2013 reach 128GB when launched? Yes.
Is that an improvement? Yes.


Could Apple add a firmware change for a CPU that is about to be widtdrawn from the Market so that "i'm never buying anything from you again" customers can squat on a 2009-2012 single Mac Pro for a longer period of time? Not very probable at all. Anyone who declares they don't want to buying anything going into the future isn't much of a customer. The machine long ago should have delivered more than it value proposition being so close to retirement. Apple isn't going to bend over backwards for those folks.

, the basis to claim a dramatically improved RAM capacity in the Tube is at best a weak ... to an outright false ... claim.

Just as weak to use Apples-to-Oranges claims.

Stick in the clarifying adjective " The new single package Mac Pro has improved RAM capacity compared to previous single package Mac Pros" and move forward.

And yes, the single CPU legacy Mac Pro is presently limited to 64GB by the use of 16GB DIMMs and will need a firmware update from Apple to use 32GB DIMMs in order to also get to 128GB....and once again, this is a technical -vs- business decision for Apple to decide *if* Apple want to bother to support legacy hardware.

Adding new specifically targeted commercial software to desupported hardware doesn't make technical or business sense. Intel is pulling the plug on the 3500/3600. This isn't an Apple only thing.

It depends on if they have 8 or 4 slots to have that option.

But what kind of user population talking about? The folks who need to jump to > 98GB in the next 2-3 years probably are in the > 64GB zone now. There are going to be extremely few folks who over the last 2-3 years though they would eventually grow to 16-64GB and now all of a sudden think that they will now need 128GB.

Even the folks who probably thought they need 64GB sooner rather than later probably jumped in at 8 DIMMs slots to use older memory tech and all 8 DIMM slots to get to 64GB faster. Just like folks who now have 64GB and want to jump to 128GB will now use older memory tech to do the same move.

Sure there is a corner case where user make major changes but that subset of of a subset of a even smaller subset isn't going to motivate a firmware change.

Of course, the other set of Mac Pro power users who have the single CPU configuration do not have this option to go beyond 64GB addressability, even though OS X currently supports up to 96GB..

Which in reality isn't a real wide spread problem.....
Even on higher geekier than average macrumors

https://forums.macrumors.com/threads/1598539/

> 64GB is down in the mid single digit percentage of Mac Pro users. If 64->80GB tapers off like 40->64GB does. ( -67% ) down to 4%. Applying same drop (no growing like rest of curve there ) the 80-128GB tapers off another -67% then down to 2.7%

While the poll here is skewed Apple probably has more accurate numbers. The new Mac Pro will cover just about everyone who is now a Mac Pro user.
The current single package Mac Pro will 90+ % of all Mac Pro users.

This whole tangent is about how Apple is highly misguided for not chasing after the last 2-3% subset of Mac Pro users. That narrow niche sub-segment doesn't make a significant difference to the long term success of the product. Not one lick. If can grow the remaining 97% by 3% they are gone with practically no impact. If can't grow that 97% by 3% then the whole product line is in trouble.

With the way it stands today, only the single CPU versions of the legacy Mac Pro (4 slots) "have to" receive a firmware update in order to install 128GB total.

Which is yet another pragmatic reason it isn't going to happen. Apple isn't going to do a "only for single package" update.

...and, as you point out, they're yesterday's old customers, not new business.

It is not so much as yesterday's customers as much as customers who primary goal is not tracking performance but to squat as long as possible on the hardware before buying anything. If Apple says the product lifecycle window is 5 years they have to stand on head and jump through hoops to get to 7-8 years just to prove Apple wrong. Some customers aren't worth having.

But yes, exchanging of money/goods/services/etc makes someone a customer. Apple pays more attention to those who are actively contributing to keeping the lights on and paying the bills.


There is a product mismatch subgroup. Apple is targeting the Mac Pro primary at folks who are looking for well rounded performance improvements over time. That is a big disconnect from folks whose workload has either plateaued or is growing much more slowly than broad spectrum tech advances.

Unfortunately, what this business case argument misses is that for some power customer use cases, a faster-Hz-despite-fewer-cores provides higher productivity than a dual-CPU system with tons of slow cores. FYI, this is particularly the case with Adobe Photoshop, as it does relatively poorly in the multithreaded arena: clock speed trumps cores.

How does the new Mac Pro miss that at all? Several of the highest affordable clock speeds in the Xeon E5 line up are single package offerings.
That is another contributing reason to why the dual package submarket got dropped.

There are 6 max now. Probably move to 8 cores at roughly same speeds with Haswell Xeon 1600 series.

Photoshop may be a RAM hog but it is not a more than 128GB RAM hog.
The 64-80GB RAM market is going to be in the "affordable" range for most of the targeted customers

But if you're going to nevertheless be optimistic, the ramifications are that technically the single CPU current Mac Pro must receive a firmware update to permit it to use 32GB DIMMs...a pretty easy litmus test.

It really isn't a litmus test of any significance. Most users are no where near the 64GB cap.

Unfortunately, this is effectively just illustrating Apple's business case which rationalized more technological stagnation than was necessary - -

It rationalized because it is rational. More expensive server/workstation folks tend to be a more risk adverse crowd. They update on slower cycles. They tend to sit and wait for 5-7 years before moving on no matter what as they wait for the ROI to pay off. they aren't buying anything so it makes zero rational sense to try to produce anything new for them. The ones that are buying are largely coming from 5-7 year old machines. A 2009 or 2010 model are both going to be much faster than a 2002-2004 machine.

It isn't just an Apple is stagnant thing. System vendors can't move too much faster than the customer. Otherwise you tend to get outbrusts just like this one around the new Mac Pro.

and yet even here things weren't completely stagnant on Apple's part because these "newer but still using the 2008 board" Mac Pro configurations were specifically obstructed from being able to use 10.6 (Snow Leopard).

Apple put in no DRM to stop folks from getting around this. 10.6 is going to be desupported probably at 10.9 transition. If putting huge hint need to move along into 2012 firmware saves Apple thousands of support calls in 2014-2016 of folks pushing 10.6 tickets because " well it was supported out of the box when I opened it" the extremely minor effort to tweak the installer is worth it.

Every other 2012 Mac has a 'Moving forward' constraint in the installer, this just keeps things consistent.


[quoet]
Understood, but we also see that other vendors were nevertheless able to provide incremental upgrades around the CPU roadmap such as SATA3 and USB3 despite Intel...[/quote]

Not really. Widespread standard configuration USB3.0 for tier 1 workstataions folks didn't come until the board redsign for Xeon E5 v1 last year. There may have been some folks who stuck PCI-e USB 3.0 cards in their BTO mazes but it wasn't standard config. Nobody did any board updates in 2011.

There aren't as many 2008 board designs for SATA 3 because SATA 3 didn't finalize till August 2008. Given the upcoming chipset completely provisioned the drive bays included Apple had little push to put a discrete SATA 3 controller in there. ( typically for core chipset it isn't until 2-3 years after standard finalization till the functionality is weaved into something everyone gets. USB 3.0 didn't make the C600 chipset that being used by Xeon E5 v1 and v2 systems now. ) Typically worktations from the 2009-2010 era that had SATA 3 tended to either have more drive bays than what the chipset comfortably supported or had a "built in RAID controller on the motherboard" feature. The Mac Pro had neither. ( and frankly most of the built-into chipset/motherboard RAID discrete chips don't work all that well well OS X anyway with the canonical drivers. )

Apple has missed SATA 3 and USB 3.0 large because they missed doing a real 2012 upgrade. That was highly likely because they were thinking about quitting altogether. (this EU market withdrawal couldn't possibly be part of any long term tactical upgrade path. At some point, Apple took a long extended detour before getting back on the road. )

 

[marshzd]

Uh...What are all the complaints about external video cards?

Tom's Hardware demonstrates there isn't much of a performance drop with video cards when used over Thunderbolt 1. And yet everyone here is saying that you only get 1/2 or 1/4 the speed, which simply isn't true.

Doesn't that mean it will be pretty much no performance drop with Thunderbolt 2? Probably a little, but really none, especially since you can aggregate...

A lot of people seem to be arguing just to argue. Here's the link to the article and breakdown:

http://www.tomshardware.com/reviews/pci-express-graphics-thunderbolt,3263.html

NOTE: I AM NO TALKING ABOUT COST, OBVIOUSLY THIS ISN'T THE BEST COST EFFECTIVE SOLUTION, I'M SOLELY ARGUING ABOUT HOW MUCH YOU CAN GET OUT OF AN EXTERNAL VIDEO CARD.

 

[beaker7]

Tom's Hardware demonstrates there isn't much of a performance drop with video cards when used over Thunderbolt 1. And yet everyone here is saying that you only get 1/2 or 1/4 the speed, which simply isn't true.

Doesn't that mean it will be pretty much no performance drop with Thunderbolt 2? Probably a little, but really none, especially since you can aggregate...

A lot of people seem to be arguing just to argue. Here's the link to the article and breakdown:

http://www.tomshardware.com/reviews/pci-express-graphics-thunderbolt,3263.html

NOTE: I AM NO TALKING ABOUT COST, OBVIOUSLY THIS ISN'T THE BEST COST EFFECTIVE SOLUTION, I'M SOLELY ARGUING ABOUT HOW MUCH YOU CAN GET OUT OF AN EXTERNAL VIDEO CARD.

Any test performed and published on Tom's Hardware is a joke. This is wideley known.

 

[marshzd]

Any test performed and published on Tom's Hardware is a joke. This is wideley known.

You say this because it completely destroys your argument? Or do you have a counter-article you could share instead of speculation?

 

[subsonix]

Any test performed and published on Tom's Hardware is a joke. This is wideley known.

It's less of a joke than people making assertive statements about it here without any testing what so ever.

 

[beaker7]

You say this because it completely destroys your argument? Or do you have a counter-article you could share instead of speculation?

What?

Did you read it?

Lets pretend for a minute that Tom's is a reputable publication.

The guy tested two obsolete graphics cards and still saw performance degradation in nearly every test, in some cases severe. It only gets worse when one puts high end cards in there.

So with external thunderbolt GPU's you get to

1) buy an external enclosure for several hundred more
2) possibly also buy another power supply as this guy had to
3) best case scenario get 90% of the performance

Or with a standard slot you can:

1) buy card, plug it in, no extra crap on your desk, get full speed.

 

[deconstruct60]

I still don't get it how Apple think they can just throw this new MP into the existing workstation environment, and expect third party suppliers to sort out the rest .
And right away, too, this clientele waits for noone to get into gear .

Three factors.

1. this clientele waits for no one after they get themselves into gear. Most workstation folks are going nowhere soon. Their Mac Pro works, this is 'new' and 'risky' and they were not planning to upgrade anyway. They'll wait for lunatic fringe to take all the arrows in back and move forward when it is a more stable solution.

2. a subset of folks have already moved on to external storage. So they they have already made open of the large moves. Long term this subgroup isn't enough a foundation for the Mac Pro, but over the first 10 month lifetime it is probably enough for a steady flow of Mac Pro sold.

3. Folks have been working on TB stuff for a long time. Belkin took almost two years to get their docking station out. If other folks started 1-1.5 years ago then later this year is about when stuff would come out of he pipeline. This "beat the farm on TB" move would have been much more risky last year because there was even less stuff out there.

Again it is how fast do folks move. Product pipelines can be 18-24 months long for something relatively completely new. Before Intel introduced the Port Ridge controller back in late 2011

http://arstechnica.com/gadgets/2011...rbolt-controller-could-broaden-reach-of-spec/

I don't think many peripherals vendors would be interested. The complexity of doing a dual port TB device is alot more than most peripheral vendors want to deal with. I'm not sure if the upcoming Falcon Ridge controllers make that easier or even more difficult with the DisplayPort v1.2 propagation and backwards compatibility mode outside of system host requirement. So most probably didn't even start seriously looking at possibilities until early 2012. Late 2013 is when stuff probably would start popping out of the pipeline.


Probably the only huge gap between folks already moving but not finished yet and this new Mac Pro is filling the zero SATA/SAS connectors inside gap the new Mac Pro introduces. I don't think anyone was trying to fill the role the standard Intel IOHUb chipset filled in covering the typical 1-4 SATA device internal to a Mac ( any of them. Even a MBA 11" uses one. Zero is extremely outside the norms. ).. The "additional and faster" storage gap, yes. That actually exists on the current Mac Pro once get to 8-18 TB storage range. There are also TB oriented devices to fill that role on the market already. The move to practically all bulk storage is outside is new and entry default capacity dropping from 1TB to perhaps 256GB is a big change.

Replacing the super cheap built in SATA controller cost is a much bigger factor. A 4-6 bay box with a low-as-possible cost single port Port Ridge controller and a very barebones SATA port controller was probably not on many folks radar. It is a bigger problem because almost no one makes 4-6 port bare bones controllers anymore because the default chipset market has wiped that out.

My guess is that someone servery under counted the number of targeted users who are using software raid (and/or ad hoc Fusion like set-ups ) using the internal drive bays and the chipsest sata controller.

 

[deconstruct60]

Tom's Hardware demonstrates there isn't much of a performance drop with video cards when used over Thunderbolt 1. And yet everyone here is saying that you only get 1/2 or 1/4 the speed, which simply isn't true.

What the speed drop off will be depends upon context.

The testing mechanism that Tom uses ( largely video games) aren't particularly sensitive to bandwidth drops. Largely because they have to deal with bandwidth drops all the time in the large number of deployed under PCI-e v2.0 lane provisioned WinPCs that are out there. The programs have algorithms to compress and cache more data on the PCI-e card so that the dramatic drop in bandwidth is nullified.

If using software that needs to move lots of data back and forth this isn't going to work as well.

Doesn't that mean it will be pretty much no performance drop with Thunderbolt 2?

No. Because the primary "savior" in the tests is the programs , not thunderbolt. As the programs expect there to be more data bandwidth ( most deployed systems moving to PCI-e v3.0 ) the "sweet spot" the programs targets is going to move to larger bandwidth requirements.

The drop will be incrementally smaller but likely reverse over the very long term as the expected amount of bandwidth goes up in newer applications.

Probably a little, but really none, especially since you can aggregate...

Depends upon where display is hooked up. If hooking the Display directly onto the external GPU then you are actually removing graphic traffic from the TB v2 backbone network. If you try to drag your eGPU's frame buffer back into the Mac and then push out to a 4K display.... LOL. it probably will get worse.


eGPU for Thunderbolt is pretty much a round peg in a square hole solution. It will work after a fashion but it is not what Thunderbolt was primarily designed to do extremely well.

 

[barmann]

Three factors.

1. this clientele waits for no one after they get themselves into gear. Most workstation folks are going nowhere soon. Their Mac Pro works, this is 'new' and 'risky' and they were not planning to upgrade anyway. They'll wait for lunatic fringe to take all the arrows in back and move forward when it is a more stable solution.

2. a subset of folks have already moved on to external storage. So they they have already made open of the large moves. Long term this subgroup isn't enough a foundation for the Mac Pro, but over the first 10 month lifetime it is probably enough for a steady flow of Mac Pro sold.

3. Folks have been working on TB stuff for a long time. Belkin took almost two years to get their docking station out. If other folks started 1-1.5 years ago then later this year is about when stuff would come out of he pipeline. This "beat the farm on TB" move would have been much more risky last year because there was even less stuff out there.

..........


My guess is that someone servery under counted the number of targeted users who are using software raid (and/or ad hoc Fusion like set-ups ) using the internal drive bays and the chipsest sata controller.

I hear you; still, what does it mean for a new MP, devoid of internal power re. storage , and no workable or affordable solution in sight ?
Storage meaning a fairly small amount of data, 1-3TB, which is cruial for performance hungry tasks.

The single, smallish, fast system drive won't do bugger all for performance, and that's all the new MP seems to come with .

Long term data storage is of course external, there is no way you could do it inside the current MP, even if you were silly enough to try it .

And of course people use software Raid for performance internally, or redundancy externally .
What the hell else , unless you have an IT guy employed ?

re 3) Much of the stuff is not out yet, hence it doesn't exist, for all intents and purposes .
What is available, is priced like it's 1999 .

The Belkin Dock is 300ish, noone knows what happens when you actually use it, and the fact remains a workstation doesn't need it, they have that stuff built in . Apart from one .

And just out of curiosity: Tell me again what TB is good for ?
I understand it might be plenty fast, like many other technologies, but it's main feature is getting rid of a seperate display port, no ?
If so, is there any benefit at all ?

 

[deconstruct60]

I hear you; still, what does it mean for a new MP, devoid of internal power re. storage , and no workable or affordable solution in sight ?

It means it makes a great SAN/NAS client. Attach a 10GbE or 8Gb/s Fiber Channel connector on it can tap into many TBs of storage that is already in place.

Storage meaning a fairly small amount of data, 1-3TB, which is cruial for performance hungry tasks.

There are probably some folks for which another 256GB or 500GB would do. It is a bit unclear if there will be an option for the second GPU card to also have a PCI-e SSD on the back. Two SSDs likely would alleviate a small amount of the storage pressure. ( especially for folks for who had already adopted PCI-e card SSD solutions in current Mac Pros for "working/scratch" space work).

The single, smallish, fast system drive won't do bugger all for performance, and that's all the new MP seems to come with .

A single drive is kind of limiting. We'll see how they price it. 512GB and 1TB are probably option.... just super high ones. Oddly enough perhaps won't look quite so high when priced again the external TB RAID box.

And of course people use software Raid for performance internally, or redundancy externally .
What the hell else , unless you have an IT guy employed ?

Some folks used RAID 0 because HDD just because of latency problems and not capacity. So they might have 3 * 256 GB dirves which present as 768 but really only needed 200GB of space. Tossing that for an SSD isn't an issue. But where speed and capacity fit in the new system makes that harder to get to.

re 3) Much of the stuff is not out yet, hence it doesn't exist, for all intents and purposes .
What is available, is priced like it's 1999 .

Not really.
http://www.blackmagicdesign.com/products/intensity/models/

Intensity Pro -- $199
Intensity Extreme -- $299 ( PCI-e card's functionality wrapped in a custom box )

The Belkin Dock is 300ish, noone knows what happens when you actually use it, and the fact remains a workstation doesn't need it, they have that stuff built in . Apart from one .

Part of that I suspect is Apple's fault too. There should be basic USB 3.0 and FW controller drivers that come from Apple that are TB aware. Apple doesn't have to support every USB and FW controller under the sun but 1-2 discrete controllers . IIRC correctly part of the Belkin hang-ups was software not hardware.

And just out of curiosity: Tell me again what TB is good for ?

Apple's docking station with the monitor built in is a pretty good fit. I suspect Apple is going to continue to be rather myopically focus it on the laptops and saddle it with a not so competitive price and size. I'm actually kind of surprised no one other vendor has introduced a serious competitor yet with better features.

Something desktop oriented without mag safe and a few more sockets ( 2 FW ports , 3 USB 3.0 , maybe multicard reader ( SD , CF , ) instead of camera , 2-3 USB 2.0 for DRM lock dongles. ) , non super glossy screen at a $500-700 price and probably have some buyers. All mac minis and Mac Pros need a screen. Align with that need. Even iMacs need a FW port at this point. LOL.

. I think the major problem is vendor don't line up the device with function. Folks may need a "sneaker net" connectivity box. ( going to get CDs , DVDs , thumb drives , USB , FW , flash cards , eSATA , TB drives coming in/out. ) So a box that handles any of that plus a temporary staging HDD inside ( so could copy and hand back a thumb drive to someone. )

I understand it might be plenty fast, like many other technologies, but it's main feature is getting rid of a seperate display port, no ?
If so, is there any benefit at all ?

It is just piggybacking on DisplayPort so that :

1. still useful if there are not TB devices in sight. (often true for first 2 years )
2. doesn't introduce "yet another port". (folks are somewhat familiar with connector. also small was an objective. )
3. the docking station/display was always an objective ( somewhat return of Apple Display Connector only suppose to be less proprietary this time. )

Thunderbolt doesn't really get rid of ports. Sometimes if the effective PCI-e card that is moved out had just one and the remote box had just one, then it disappear ( e.g., low end eSATA card and single socket eSATA enclosure ).


What Thunderbolt is sorely missing is :

1. Affordable fiber. If there is a distance difference (put my computer in the next room ) component some folks might find that attractive.

Affordable fiber also means would eventually be able to go faster.

2. More integration. For instance Intel combines part of its core chipset blocks for USB 3.0 ( 3-4 ports) and SATA (1-2 6Gbp/s ports 1 3Gb/s port ) , 1GbE and perhaps HD Audio into a TB controller that specifically targets docking stations that also fully handles backward compatible DisplayPort mode.
In other words take most of what is in a belkin like like docking and integrate it into a one chip solution that folks can wire up. maybe leave x1 PCI-e lane coming out so that some can add small differentation ( another USB controller or FW controller or whatever) but most of it is ready to go out of the box.

20 vendors all using the same chip and its associated R&D is going to drive down costs.

Intel could do another variant that bundles their SATA/RAID block with the controller.

 

[-hh]

I loop back to 5 years in after clearing up "support"

Err, no. This has to do disconnect between what provide support services technically and the notion of "support"/"supported" meaning "being made to work somehow".

Understood, but my point was merely that there's also consumer expectations on what 'support' may electively entail, and while Apple's specific obligations may be limited, they can exceed them at their discretion.

Technically Apple's policy is 5 years after discontinuation. This super long extended transition and shared motherboard over full tick-tock cycle pragmatically presents a big disconnect here. The 2010 (model 5,1) technically could start the clock on the 2009 ( model 4,1). Many of the single package Mac Pro even into 2012 variants are 2009 era CPUs.

The 2010 (5,1) did start the 5 year clock on the 2009 (4,1), as did also the 2012 (6,1) on the 2010. And you're correct that this is not a technology constraint, but rather a business (support policy) decision.

Intel is actually going to pull the Xeon 3500/3600 from the retail market in September 2013. They are going to pull them from system vendors in 2014. Apple's 5 year window is not something they just pulled out of their butt...

True, but you've really missed the important point: Intel's lifecycle roadmap reveals that Apple couldn't continue to maintain the status quo of the aging Mac Pro motherboard -- it (the motherboard) was coming up onto a "Kill It -or- Change It" constraint, and Apple had to decide if they were going to abandon the Mac Pro, or make a reinvestment in a new design & tooling.

To be blunt there s a bit of disingenuous spin here on both sides. The reality is that the dual package Mac Pro is dead. If want to an Apples-to-Apples comparison of new Mac Pro to old Mac Pro, then it is really single package variant versus single package variant.

Sorry, but your claim conflates *features* for *capability*.

The customer requirement is for performance (capability) - they objectively should not care how a capability is achieved inside of the proverbial black box ... it notionally could be one CPU with 8 cores, or it could be 2 CPUs of 4 each, or even 4 CPUs of 2 each, etc ...

The current Mac Pro is sold at a couple of basic capability levels...we do conflate these with specific features, such as CPU count, #cores, RAM slots, etc, because features do generally correlate to capability - but there are exceptions such that we cannot blindly rely on features in lieu of capability.

Could Apple add a firmware change for a CPU that is about to be widtdrawn from the Market ... Not very probable at all.

Which is what I said. Thank-you for agreeing with me.

the basis to claim a dramatically improved RAM capacity in the Tube is at best a weak ... to an outright false ... claim.

Just as weak to use Apples-to-Oranges claims.

Stick in the clarifying adjective " The new single package Mac Pro has improved RAM capacity compared to previous single package Mac Pros" and move forward.

Sorry, but no: the addition of your qualifier (single CPU) is a technically unnecessary, artificial and utterly contrived constraint.

Particularly since the fact remains that a current Mac Pro is still being sold by Apple which is a configuration which supports 128GB of RAM, which is the same max RAM capacity that is being advertised for the new Tube Mac Pro (4x32).

The only legitimate claim that you can make is that Apple doesn't sell an OEM configuration with 8x16GB...but that's being very shortsightedly pedantic, as that's a business decision, not a technological limitation.

Adding new specifically targeted commercial software to desupported hardware doesn't make technical or business sense. Intel is pulling the plug on the 3500/3600. This isn't an Apple only thing.

Except that there's already commercial software which today is being employed byMac Pro's equipped with greater than the 64GB "limit" of Apple's OEM configuration. Plus, that's really a business decision for those 3rd party vendors to decide, particularly since all of the pre-2013 Mac Pro's aren't going to suddenly disappear...there's an installed base of consumers to consider.

But what kind of user population talking about?[/quotes]

Ones with capability needs who are more willing to fork over the dollars to resolve their workflow issues than the casual recreational hobbyists and fanboys.

The folks who need to jump to > 98GB in the next 2-3 years probably are in the > 64GB zone now.

Yes, and technologically, we should note that the 16GB DIMMs they used is a non-Apple-OEM part. As soon as the 98GB limit is raised by 10.9, they'll probably perform another incremental upgrade their current Mac Pro hardware.

There are going to be extremely few folks...

The chronic issue for the Mac Pro has precisely been that it is sold to extremely few ... and the ones who buy the more capable dual-CPU version are fewer still.

The real ramifications of the Tube Mac Pro is to cut Apple's fixed cost of development, production, logistics and support from two discrete models (single & dual CPUs) to one. That business case approach is very consistent with the ongoing decline in the sales of IT "Trucks" at Apple.

As such, the only real question is if this single configuration is able to address the diversity of current Mac Pro consumer needs and if not, who is being abandoned in order to realize this cost savings cut.

The new Mac Pro will cover just about everyone who is now a Mac Pro user.

I've heard such promises made before.

In any case, the real question isn't necessarily "capability", but the cost to field that capability.

For example, given the current (2013) state of the shelf, to provision four 3.5" hard drives via Thunderbolt instead of internal bays is by my rough estimates approximately a $600 higher expense. As such, it won't be known until Apple releases specifics with prices for us to know if the value paradigm for the new Mac Pro is better or worse than the current.

This whole tangent is about how Apple is highly misguided for not chasing after the last 2-3% subset of Mac Pro users.

Perhaps for some, but I was motivated to correct the factually incorrect claims, flawed logic and similarly incorrect assumptions being made.

It is not so much as yesterday's customers as much as customers who primary goal is not tracking performance but to squat as long as possible on the hardware before buying anything. If Apple says the product lifecycle window is 5 years they have to stand on head and jump through hoops to get to 7-8 years just to prove Apple wrong. Some customers aren't worth having.

Spoken like a businessman who hates his customers ... and is it merely because they're value-concious as opposed to fickle, shallow consumers?

Now this does bring up a very different and far reaching tangential question: to what degree has Apple shifted their priorities into businesses which are more volatile and subject to the whims of consumer fads? Before you try to dismiss this question, do consider the history of Sony, as they were previously one of the "can do no wrong" darlings of the consumer marketplace...Trinitron, Walkman, others...are all gone today.

There is a product mismatch subgroup. Apple is targeting the Mac Pro primary at folks who are looking for well rounded performance improvements over time. That is a big disconnect from folks whose workload has either plateaued or is growing much more slowly than broad spectrum tech advances.

An interesting declaration, given that we were just told that Apple is highly misguided....

RE: 1x6core vs 2x4

How does the new Mac Pro miss that at all? Several of the highest affordable clock speeds in the Xeon E5 line up are single package offerings. That is another contributing reason to why the dual package submarket got dropped.

Sorry, but I'm simply reporting the facts on the current generation...whereas it appears that you're speculating on what the future Mac Pro might be sold with.

And while you can point to the Intel roadmaps for the basis of your speculation, the problem with that approach is that it ignores the history that much of the advanced Mac Pro consumer group's frustration has been with this same type of knowledge from Intel yet which does not get applied into products available from Apple...particuarly if we add an affordability metric - - just witness all of the DIY CPU upgrade threads on MR as examples.

There are 6 max now. Probably move to 8 cores at roughly same speeds with Haswell Xeon 1600 series.

Keyword being "Probably" . Those that really rely on a Mac Pro for their livelihood won't buy on Day 1, but will wait until there's some benchmarks out so as to do a Cost:Benefit analysis.

Photoshop may be a RAM hog but it is not a more than 128GB RAM hog...

Point missed yet again: it is the question of using Photoshop in the 98 to 128GB range.

The 64-80GB RAM market is going to be in the "affordable" range for most of the targeted customers

Actually, for some of these higher amounts of RAM, it is more affordable to install in a 2012 dual CPU Mac Pro than in the new 2013 Tube, for the very reason that 2x16 is less expensive than 1x32 and these machines have 8 slots instead of 4.

It really isn't a litmus test of any significance. Most users are no where near the 64GB cap.

Actually, "Most Users" (80%) are using laptops...but then again, we do need to better understand the target customers for any particular solution before daring to make generalizations.

It isn't just an Apple is stagnant thing. System vendors can't move too much faster than the customer. Otherwise you tend to get outbrusts just like this one around the new Mac Pro.

Actually, Apple has already screwed over some of their Enterprise professional customer base because they keep on updating OS X faster than it can get through vetting for approval for use...in this regards, the un-termination of Snow Leopard was a minor godsend - - except that the 2012 Mac Pro is blocked from reverting to it.

All in all, this is one of the reasons why Apple does poorly outside of consumer sales ... and I see that the same is already happening with iOS too: one large Enterprise just approved iOS 6 for deployment, so if Apple stops selling it with iOS 7 launch, they just lost an iPhone customer who is looking at replacing 200,000 RIM Blackberries.

Understood, but we also see that other vendors were nevertheless able to provide incremental upgrades around the CPU roadmap such as SATA3 and USB3 despite Intel...

Not really. Widespread standard configuration...

Nobody did any board updates in 2011....there aren't as many 2008 board designs...

Seems that we're missing declarative statements for 2009 and 2010...and quite frankly, even 2012 as well.

Apple has missed SATA 3 and USB 3.0 large because they missed doing a real 2012 upgrade. That was highly likely because they were thinking about quitting altogether. (this EU market withdrawal couldn't possibly be part of any long term tactical upgrade path. At some point, Apple took a long extended detour before getting back on the road. )

Agreed. However, what's supremely frustrating is that Apple couldn't have even bothered to have offered an OEM-supported PCIe card with USB3.

...
My guess is that someone servery under counted the number of targeted users who are using software raid (and/or ad hoc Fusion like set-ups ) using the internal drive bays and the chipsest sata controller.

More likely IMO is that Apple looked at only the numbers of what they've sold...

...and despite the fact that their in-house Crash Log Reports will tell them precisely all of the additional upgrades (3rd Party) had been provisioned, they probably never even looked. Afterall, that's "Tech Support", not "Sales" and a classical stovepipe organizational failing.


-hh

 

[thekev]

Photoshop may be a RAM hog but it is not a more than 128GB RAM hog.
The 64-80GB RAM market is going to be in the "affordable" range for most of the targeted customers

Photoshop took a custom virtual memory implementation rather than a lot of additional files, but the more demanding cases of large plates with 100+ layers (which is partially due to limitations in layer structure) existed in the Power PC era too when it was still a 32 bit app. The data that needed to be moved around was still quite large. It's just now it tends to handle more directly in ram rather than by aggressively moving it to disk. It can still grow with time, but the ram hunger is going to be fairly logarithmic unless you're dealing with ever increasing resolution. With composite images that is unlikely. Just cleaning up a file from even bleeding edge 80MP digital backs won't generate that amount of data. If you're working with that much regularly, it's understandable to need 32. Beyond that you're likely to hit heavy diminishing returns. With the Digilloyd stuff he's often testing 15k resolution and beyond, which places it back in the realm of the heavy comp work I mentioned.

I think it's not always an issue of what people need today but a concern that it doesn't leave enough headroom for later. Four dimms are likely to be populated initially by fairly demanding users, so anything more requires removing old sticks. It adds a bit to the price, but I suspect that not many of these users will need to go past 64GB with this machine. If 16GB dimms get down to a lesser premium by next year, I think many people will just go with that for the next 3-4 years. Software developers tend to tune somewhat for what their users have available.

 

[shaown]

Radiating - I just went through this whole thread. While I'm pretty excited about the new Mac Pro, you (and a few others) seem to have lost sight of what should get compared. I keep seeing references to how the new 2013 Mac Pro is better than (slightly) or not quite as good or whatever versus the 2010 Model. Cmon folks, while some people have to run OSX, most workstation users (especially groups buying 10+) will consider other platforms. I hope to god a 2013 model can match up to and exceed the 2010 for performance. The only valid comparison is how will the 2013 Mac Pro stand up to the latest workstation class machines running Linux, Windows, etc. And I think the conclusion is pretty obvious, if priced right, the new machine will compete well / exceed SP machines. Especially with all its built ins (PCIe SSDs, etc). But for raw power, it won't compete with machines that are DP 12 core machines (i.e. 24 physical cores) and take up 8 x 32 GB DIMMs (or 8 x 16 at a lower price).
-Shaown

MacVidCards, your repeated personal attacks and fixation on argumentative nonsense indicate a profound disinterest in being impartial and rational about your views. You are someone who clearly starts with a viewpoint and then tries to prove it instead of the other way around.

 

[pearvsapple]

Still functions as any desktop form factor out there. In the end, proprietary components and fugly chassis design. This is the Steve Jobs difference.

 

[Snowshiro]

Maybe OP is right. Maybe I don't understand the new Mac Pro.

But I understand this:

As a Mac Pro owner looking to upgrade my workstation in the next cycle, Apple told be to hold on because they had something good in the pipeline.

Now it's been revealed, I'm sorry I waited. As a pro user with a need for a high performance workstation, I'm sick of Apple trying to tell me what I want rather than listening. Final Cut and now this means I'll be going Windows.

 

[calaverasgrande]

The chronic issue for the Mac Pro has precisely been that it is sold to extremely few ... and the ones who buy the more capable dual-CPU version are fewer still...

For example, given the current (2013) state of the shelf, to provision four 3.5" hard drives via Thunderbolt instead of internal bays is by my rough estimates approximately a $600 higher expense. As such, it won't be known until Apple releases specifics with prices for us to know if the value paradigm for the new Mac Pro is better or worse than the current.

In my workplace, every single Mac Pro is dual CPU. I don't know who buys the single processor boxes?
Second, I think that is a great point. We can't really say which way we will go on the next iteration of Mac Pros until we know how much it will cost to buy the max spec version with an external PCIe box. We generally do not sweat workstation prices, but if we can hop over to a Z820 with dual Xeons for the price of a Mac Pro 2013 + PCIe box, we will.

 

[beaker7]

In my workplace, every single Mac Pro is dual CPU. I don't know who buys the single processor boxes?
Second, I think that is a great point. We can't really say which way we will go on the next iteration of Mac Pros until we know how much it will cost to buy the max spec version with an external PCIe box. We generally do not sweat workstation prices, but if we can hop over to a Z820 with dual Xeons for the price of a Mac Pro 2013 + PCIe box, we will.

+1.

I've never seen a single CPU Mac Pro in person.

Have deployed probably 100 duals over the past 7-8 years

 

[zephonic]

In my workplace, every single Mac Pro is dual CPU. I don't know who buys the single processor boxes?
Second, I think that is a great point. We can't really say which way we will go on the next iteration of Mac Pros until we know how much it will cost to buy the max spec version with an external PCIe box. We generally do not sweat workstation prices, but if we can hop over to a Z820 with dual Xeons for the price of a Mac Pro 2013 + PCIe box, we will.

You will see more single-CPU models in music and audio, as many audio applications don't really benefit from the dual CPU anyway.

A single W3680 will generally be better suited for music stuff than a dual quadcore.

 

[slughead]

Tom's Hardware demonstrates there isn't much of a performance drop with video cards when used over Thunderbolt 1. And yet everyone here is saying that you only get 1/2 or 1/4 the speed, which simply isn't true.

Uh that toms hardware article 1) Used old video cards 2) Specifically SAID that faster video cards would suffer.

At least you don't have to worry too much about procuring a flagship graphics card, right?[...] Pinching off that bus means that, at a certain point, it doesn't matter how big of a GPU you attach to the host.

http://www.tomshardware.com/reviews/pci-express-graphics-thunderbolt,3263-8.html


Other benchmarks show that 600 series+ NVidia cards suffer immensely when bottlenecked through PCIe 2.0 4x (2GBps). Radeons seem to be bottlenecked less (funny, considering the 7970 was touted as the first PCIe 3.0 16x card)

Try putting a Titan or GTX690 into one of those little chassis. Even with TB2 it'll be a huge performance hit.


TB is a GREAT interface, but it cannot hold a candle to PCIe for graphics, nor was it designed to. The whole idea is ridiculous.

 

[ness96]

In my workplace, every single Mac Pro is dual CPU. I don't know who buys the single processor boxes?
Second, I think that is a great point. We can't really say which way we will go on the next iteration of Mac Pros until we know how much it will cost to buy the max spec version with an external PCIe box. We generally do not sweat workstation prices, but if we can hop over to a Z820 with dual Xeons for the price of a Mac Pro 2013 + PCIe box, we will.

I use a single quad core 5,1. I work as a software engineer. I mainly like the Pro's power for virtual machines, and the quad core more than covers this. Of course, I don't think many Pro owners use the machines for this purpose.

True to your point, from a media editing standpoint, I don't understand why anyone would buy a quad core Pro.

 

[liquid stereo]

wrong observation

I think the more pertinent/relevant observation would have been "I've never seen low-end Mac Pros in person".

Working professionals tend to get the highest performance spec they can afford.

Cheers!

+1.

I've never seen a single CPU Mac Pro in person.

Have deployed probably 100 duals over the past 7-8 years

 

[fastlanephil]

Radiating posted this.

Or maybe in something like this someday.

 

[deconstruct60]

Radiating posted this.

Probably a very quick way of screwing up the new Mac Pro. Put it in a box so the vents at bottom can ingest some of the air coming out of the top and create a feedback loop.

Image

 

[deconstruct60]

In my workplace, every single Mac Pro is dual CPU. I don't know who buys the single processor boxes?
S.

+1.

I've never seen a single CPU Mac Pro in person.

That isn't mainstream Mac Pro user sampling.

Have deployed probably 100 duals over the past 7-8 years

Let's say Apple sold 30K Mac Pros per year over last 7-8 years. (pretty dismally low but being conservative ) So that is 210K to 240K. 100/210K 0.05% .

the 100 sounds like alot until look at how many Mac Pros are out there. Nor is working in a highly vertical industry a broad spectrum demographic sampling methodology.

Finally there is also the fact that you had to go dual package to get to 2 and later 4 cores if go back far enough in time. What really talking about then was being able to cross the 2 core mark that every Mac in the line-up reached or blew past years ago. In another year (maybe two ) every Mac will probably reach the 4 core minimum. So really not saying much of anything about the good old days when needed two just to go multiple.