Is the new Mac Pro a failure?

[thefredelement]

It's not a failure, it's not defective in anyway and does exactly what it was designed to do. It's designed to kick butt using Apple's pro apps and it does that. It wasn't designed for CUDA computation, it's not advertised with 6TB of internal storage.

It took almost 6 months for it to catch up with demand, that is a very dramatic statement considering how well Apple delivers inventory.

I think the real failure, that people then use to perceive the 2013 MP as a failure, was the lapse in pro software by Apple and maybe not releasing the 2013 MP form factor sooner. If they didn't let FCP7 stagnate for so long and then drop a dud with FCPX but instead released a good FCPX alongside a 2013 style MP in 2011 or 2012 this thread wouldn't exist.

If Apple cared about altering their platform to support Premier they would have done so, but they designed metal instead and one day that's something Adobe is going to have to embrace if they want to deliver on OS X like they do on Windows.

You can't get a FCPX workstation to compete with the nMP without significant modifications. Modifications mean more redundancy, more points of failure and more stuff to just have to setup and worry about.

The 2013 MP is the best workstation if you're a pro using Apple's pro apps. If you make money using it and it saves you time or aggravation (in rendering or compilation or warranty repair, really in anyway) then it is a success. If it's performance allows you to recover your investment into the hardware over and over again then it's a great success and you have used your tool for work as designed.

If you're sad because you can't have an internal RAID array and swap out video cards every time someone iterates then maybe the Mac Pro is no longer for you, a PC workstation sounds like it would work better for you. Seriously, if you want to geek out about PCI slots then do it, buy all the older MPs or PC workstations or whatever has PCI slots and just geek out over them.

If I need a hammer drill I like to use my Bosch hammer drill, I don't like to use my cordless that's also a hammer drill, maybe it can get the job done here and there on the fly but if I need to get through a real amount of concrete I'm not going to mess around with the cordless. My Bosch hammer drill is probably around 15 years old now (maybe even older), there are new cordless hammer drills released all the time. They're shiny, they probably do better in benchmarks but there's no way they're as adapt at performing the same type of work a dedicated hammer drill is. There are new versions of this hammer drill and when the one I have no longer gets the job done fast enough for me to be successful then I will get a new version of the best tool for the job.

This thread is a moot point, the e5 v4 (or really any future Xeon version) if/when released in a MP, will obliterate any previous Mac in real world performance and benchmarks. There will only be conformance to whatever that form factor supports if that's the best tool for your job.

This sub forum on MR seemed to be an oasis of good Apple Pro geeky stuff that felt far away from the "it's too heavy, it's too big, it's too small, I want an android, there's a 0.00003838292" part of my screen that has a slight yellow tint" threads and I really hope it gets back there without people insulting each other.

 

[iSee]

What made the cMP a failure?

Only Apple knows and they probably won't say.

In the past Apple has said they only want to enter markets if they have something significant and unique to add... That's probably the same criteria they use to decide whether to stay in a market or not.

So perhaps they felt they had little to offer in a traditional workstation. That is, if the cMP was going to be essentially identical to the workstations HP, etc. were selling, then why bother?

 

[H2SO4]

It's not a failure, it's not defective in anyway and does exactly what it was designed to do. It's designed to kick butt using Apple's pro apps and it does that. It wasn't designed for CUDA computation, it's not advertised with 6TB of internal storage.

The fact that its remit is so narrow to me makes it a failure. If it was designed to kick but with Pro Apps in general then yep, it’d be good but for Apples pro apps only. Crap.
That’s like saying an F1 car designed to be good on street circuits where most circuits are in fact not, is a towering achievement. A good all rounder can perform brilliantly everywhere.

 

[rawweb]

Only Apple knows and they probably won't say.

In the past Apple has said they only want to enter markets if they have something significant and unique to add... That's probably the same criteria they use to decide whether to stay in a market or not.

So perhaps they felt they had little to offer in a traditional workstation. That is, if the cMP was going to be essentially identical to the workstations HP, etc. were selling, then why bother?

Can you possibly cite the source for that statement? I'm curious what context that might have been in. I have a hard time believing that Apple only enters markets unless they have something significant to add to it. They are after all a company trying to make money. If there's a dollar to be had, I'm sure Tim Cook would build it. I also can't get behind the theory in your final statement. If that was the case, they would have left most of the markets they currently serve. I think it's even simpler than that: Mac Pro needed to be brought in line with the rest of the product lineup. An era where upgradability means a stick of ram, small form, little noise and low energy consumption (and much less on site retail genius training).

I use a 6,1 for 4k video production and web design at work. The hardest (and most annoying, expensive etc) transition was different/new storage solutions. But you know what? It gets the job done, really well. So, based on the criteria of this poll, it is not a failure for me. I do about the same kind of work on the side at home on a 5,1. It stacks out about the same in my eyes. For the work I perform, they both are phenomenal machines. If I could afford one at home, would I buy it? I'm afraid to say on this forum.

Ever since I first unboxed the computer I always imagined audio engineers would absolutely love the thing.

 

[thefredelement]

The fact that its remit is so narrow to me makes it a failure. If it was designed to kick but with Pro Apps in general then yep, it’d be good but for Apples pro apps only. Crap.
That’s like saying an F1 car designed to be good on street circuits where most circuits are in fact not, is a towering achievement. A good all rounder can perform brilliantly everywhere.

You're literally defining a PC workstation, not a MAC.

 

[MacVidCards]

since just about every thread i participate in has people saying 'you bought an imac' in some form or another, typically in derogatory fashion.. i'm going to lay out the situation once and for all so i can link back to this post next time it happens.. it's not meant as some sort of point re:this particular discussion so skip over if desired:

for starters/background, 'my computer' is not a single item.. it's all the stuff related to computing that i use.. in 2013, i had a 1,1 and a 2010 core2 mbp.. my main application, the heart of my computing needs, is a 3d modeling/cad application(s).. CAD relies heavily on single core performance as most(!) of the mathematics involved are linear calculations and simply can not be spread out amongst multiple threads/cores.. for sales purposes however, i rely heavily upon ray traced renderings which benefits greatly from multicore processors..

i planned on buying a 6core/d700 config.. i had a 6month warning from my cad devs (mcneel) that they were going to be dropping support for OS X 10.7 and earlier (and the 1,1 doesn't support anything after 10.7)..
when the latest mac pro was announced, my rendering software developers (glare tech.) announced they were planning to completely re-write indigo using openCL at the core.. from a mac user pov, this can be interpreted as "apple says nmp is capable of up to 7tflops processing power.. well, we're going to give you the keys to unlock that potential"
so yeah, that's awesome news for me.. a 6core with d700 is going to out perform a 100+ cpu core renderfarm.. great.
except when august2014 rolls around and my cad software is weeks away from dropping 1,1 support (vis a vis apple).. i can't hold out any longer.. i must buy a new computer now which will run the latest versions of rhino.. unfortunately, indigo (renderer) was still only in it's early phases of the openCL rewrite.

that's the situation i was in.

what i decided to do.. what i truly feel was the best course of action.. was this:

buy a refurb iMac (RFB IMAC 27"/3.5QC/8GB/3TB-FD/780M-USA).. give it 16 more gb ram.. (about $2500)... to hold me over for a year or two

but-- there's more to it than just that.. the iMac isn't 'my computer'

using the money budgeted for the 6.1 and a new display-- use the rest like:
• early replace the mbp to a quad with 802.11ac wifi. (wasn't due for replacement till 2016)
• obtain larger cloud drive (well, 2.. dropbox and iCloud)
• upgrade my internet speed (200mbps)
• upgrade my router to 802.11ac
• buy wifi drives and move mass storage and non-cloud backup to them.


taking advantage of OS X's updated networking capabilities along with some improvements from glare.. and utilizing the improved speed of wifi.. it's incredibly simple for me to use the iMac and mbp as a single unit when rendering..
yes, they throttle upon prolonged usage but i have 8 cores instead of the 6 which were going to be in my 6,1.. i haven't nerd tested it but i'm guessing the 8cores throttled down are giving me something close to what i'd expect from the 6 cores of a 6,1..

my renderer couldn't yet utilize the d700s so they'd more or less just sit idle during the times i'd need them the most.. it was still a choice of multicore cpu usage at that time..

i spent approx equal amount of money on my current set up as i would of with the 6,1 and i'm getting approx equal performance..

the advantages of what i did however (at least as i see it) is that my infrastructure is modernized.. my laptop has been updated to a point where i can now do smaller renders during lunch break where as the 2010 just wouldn't cut it..

when the 7,1 releases.. and indigo4 releases.. then it will be time to sell the iMac and slip the mac pro back into the setup.. i'll have (i assume) better than d700s in it which will even further utilize the updated software.. moreso than if i had d700s sitting idle for the past two years.. the d700s will still run indigo great (and is doing so in the developer previews of the software).. it's just that 7,1 will theoretically do even better.

but there was no way i was going to buy a 6,1.. then a 7,1.. which would also coincide with the scheduled update of the 2010mbp..

that's just a poor plan and not the best use of computer budget.. i don't know- i really do feel like i made smart and well informed decisions when bypassing the 6,1 but apparently, it's something i should be ridiculed over around here.









yes, without a single doubt, those people are here.

Why? well, last time i started talking MR_member psychiatry around here, i got a timeout and the posts were deleted.. i'll respect the forum rules regarding this type of talk and not get into it again. ; )

Wow, there is a wall of words.

So, are you saying that you needed CUDA and nMP had none so you had to buy a machine with an Nvidia card? Oh the irony...

So while the 6,1 failed to meet your needs, a different computer met them so you bought that instead, thus voting with your wallet?

And where are those nMP GPU upgrades you predicted? Pretty sure you said 12-18 months a couple years back. Are you still post dating those or finally ready to admit that they are officially "vapor-ware"?

 

[-hh]

A couple of observations (I'm only on page 7 so far of this thread).

First, the question of "what constitutes success/failure?" is a very astute question. From the perspective of Apple as a business, they're probably achieving 'success' by however they define it ... but that's not the same metric as to what their customers think.

I noticed a couple of comments which mentioned respective capabilities of the nMP versus cMP. One was hardware (number of RAM slots, etc) which was pedantically correct, but tip-toed around the real issue. The other hit the nail square on: it isn't if "Capability X" can't be cobbled together by some means (TB-PCIe adaptor box-Card, etc), but really if that all works out to be as good of a value.

For example, compare & contrast the cost of a couple of bare internal drives thrown into a cMP versus buying a Thunderbolt Promise Pegasus R4 external ... the latter duplicates the capability, yes, but does so at a substantially higher cost. As such, we can clearly conclude that the nMP is a 'failure' for those use cases who require such storage because it has indisputably regressed on the value curve (at very best, it is a zero net cost change "push" for those users who wanted it external anyway for reasons of use case portability).

And what this is really alluding to is that while the cMP was a generalized "truck" which a broad array of Pros/PowerUsers would then tailor to their specific needs, the nMP is clearly a specialized piece of kit which does very poorly at certain types of tailoring.

Second, I saw some disagreement over the Pro vs Power User customers as a source of opinions ... the short answer here really is "But Who Cares?" Well, Apple has to (theoretically), because they want people to buy their stuff. And what this really means is that unless we know for ourselves just what the "Power" versus "Pro" demographic sales distribution statistics are, there isn't any valid means by which we can disregard one group over the other...and the broader implications of this are that Apple (and us) had better then strive to listen to the inputs of both, if the broad strategic objectives are to (a) retain them as customers and (b) serve both sub-markets with a single hardware configuration (streamlined manufacturing)

Finally, there's this:

Then I would recommend you start investigating a migration away from the Mac soon. Making a move can be time consuming so better to start seeing what's available and what's involved before you're forced to. If I were a professional who relies on Mac products for a living I'd be doing so. Apple has signaled they don't care about professionals any longer.

Where Apple has been successful with the Mac over the past decade is in product differentiation - - which is also why they're able to clear 30%+ profit margins, unlike the Windows PC razor-thin "Commodity" hardware market. But the means by which they did much of this was through productivity applications - - we don't run a computer to just stare at the OS. As such, the question is if Apple has been continuing to cultivate these ecosystems, or have let them wither and die.

Final Cut ... going from Pro to X has had a 60% decline in its marketshare since 2007, despite a huge price cut.
Aperture ... abandoned.
iPhoto .. extremely dumbed-down.

And so on.
Apple is reaping what they chose to not sow.

In the meantime, the Magic Mouse 2 has had a crackerjack engineering team to fix a merely bad-sounding 'click'.


-hh

 

[iSee]

Can you possibly cite the source for that statement? I'm curious what context that might have been in. I have a hard time believing that Apple only enters markets unless they have something significant to add to it. They are after all a company trying to make money. If there's a dollar to be had, I'm sure Tim Cook would build it. I also can't get behind the theory in your final statement. If that was the case, they would have left most of the markets they currently serve. I think it's even simpler than that: Mac Pro needed to be brought in line with the rest of the product lineup. An era where upgradability means a stick of ram, small form, little noise and low energy consumption (and much less on site retail genius training).

I use a 6,1 for 4k video production and web design at work. The hardest (and most annoying, expensive etc) transition was different/new storage solutions. But you know what? It gets the job done, really well. So, based on the criteria of this poll, it is not a failure for me. I do about the same kind of work on the side at home on a 5,1. It stacks out about the same in my eyes. For the work I perform, they both are phenomenal machines. If I could afford one at home, would I buy it? I'm afraid to say on this forum.

Ever since I first unboxed the computer I always imagined audio engineers would absolutely love the thing.

Here's an article with the quote:
http://www.standard.co.uk/lifestyle/london-life/sir-jonathan-ive-the-iman-cometh-7562170.html

Our goals are very simple - to design and make better products. If we can’t make something that is better, we won’t do it.

There might have been some Tim Cook quotes along the same lines.

Apple is a company trying to make money, but their strategy is not to turn a dime on any little thing they think they can turn a dime on, but to focus on products where they can make a it better.

So, perhaps Apple walked away from traditional workstations (ones like the old Mac Pro) because they didn't think they could make them significantly better than others.

Like I said before, I don't know why Apple dropped the old Mac Pro. I'm only speculating.
They could have easily kept updating it -- they have all the money, expertise and other resources they would need to do that. They could have even kept it going along side the new Mac Pro if they wanted to. But Apple chose not to. I think that's significant and means it's not coming back, regardless of what happens with the new Mac Pro.

 

[thefredelement]

A couple of observations (I'm only on page 7 so far of this thread).

First, the question of "what constitutes success/failure?" is a very astute question. From the perspective of Apple as a business, they're probably achieving 'success' by however they define it ... but that's not the same metric as to what their customers think.

I noticed a couple of comments which mentioned respective capabilities of the nMP versus cMP. One was hardware (number of RAM slots, etc) which was pedantically correct, but tip-toed around the real issue. The other hit the nail square on: it isn't if "Capability X" can't be cobbled together by some means (TB-PCIe adaptor box-Card, etc), but really if that all works out to be as good of a value.

For example, compare & contrast the cost of a couple of bare internal drives thrown into a cMP versus buying a Thunderbolt Promise Pegasus R4 external ... the latter duplicates the capability, yes, but does so at a substantially higher cost. As such, we can clearly conclude that the nMP is a 'failure' for those use cases who require such storage because it has indisputably regressed on the value curve (at very best, it is a zero net cost change "push" for those users who wanted it external anyway for reasons of use case portability).

And what this is really alluding to is that while the cMP was a generalized "truck" which a broad array of Pros/PowerUsers would then tailor to their specific needs, the nMP is clearly a specialized piece of kit which does very poorly at certain types of tailoring.

Second, I saw some disagreement over the Pro vs Power User customers as a source of opinions ... the short answer here really is "But Who Cares?" Well, Apple has to (theoretically), because they want people to buy their stuff. And what this really means is that unless we know for ourselves just what the "Power" versus "Pro" demographic sales distribution statistics are, there isn't any valid means by which we can disregard one group over the other...and the broader implications of this are that Apple (and us) had better then strive to listen to the inputs of both, if the broad strategic objectives are to (a) retain them as customers and (b) serve both sub-markets with a single hardware configuration (streamlined manufacturing)

Finally, there's this:



Where Apple has been successful with the Mac over the past decade is in product differentiation - - which is also why they're able to clear 30%+ profit margins, unlike the Windows PC razor-thin "Commodity" hardware market. But the means by which they did much of this was through productivity applications - - we don't run a computer to just stare at the OS. As such, the question is if Apple has been continuing to cultivate these ecosystems, or have let them wither and die.

Final Cut ... going from Pro to X has had a 60% decline in its marketshare since 2007, despite a huge price cut.
Aperture ... abandoned.
iPhoto .. extremely dumbed-down.

And so on.
Apple is reaping what they chose to not sow.

In the meantime, the Magic Mouse 2 has had a crackerjack engineering team to fix a merely bad-sounding 'click'.


-hh

I think there needs to be some counter observations to your observations, one thunderbolt2 is substantially faster than the sata2 interface found inside the classic Mac Pro tower.

Second the Photos app has had a huge amount of development in an effort to replace iPhoto at an OS level. Some could argue that the new interface and integration (across all of Apple's platform) offers a vastly more unified, both in look and function experience. I would happen to agree with that sentiment.

I 100% agree with your Final Cut observation and think the 2013 Mac Pro along with FCPX.1+ is Apple attempting to reclaim some of that user base. I think it's an uphill battle for them but as I've just posted in a different thread I think Apple is a unique position to deliver next gen video better than others can given their advancements on mobile and further cross platform development of the Metal API.

 

[-hh]

I think there needs to be some counter observations to your observations, one thunderbolt2 is substantially faster than the sata2 interface found inside the classic Mac Pro tower.

True enough on the surface - - but:

First, SATA technology moved to SATA-III (6Gbps) circa 2008/09m, so not only could something faster than SATA-II been adopted in the 2010 Mac Pro update, it became downright shameful that it wasn't incorporated in the 2012. As such, we're differentiating based on false pretenses.

Second, performing incremental updates (vs wholesale system replacement) is precisely what the appeal is for things like PCIe slots in towers. Case in point: what provisions exist in the nMP to upgrade a to a higher performance subsystems? Granted, we can debate about what percent of customers do upgrades, but when there's zero provisions to upgrade at all, this too becomes a moot point.

Third, Thunderbolt certainly has speed, but it also has latency which makes it a not-quite-equal to native PCIe. Which may or may not matter for one's use case, but needs to be noted to see if it is an important differentiator or safely irrelevant.

Fourth, the performance bottleneck on traditional spinning drives is the drive itself, not the interface. As such, SATA-III is primarily only important if one wants several SSDs in drive bays so as to not give up 'too many" PCIe slots for the capability.

Second the Photos app has had a huge amount of development in an effort to replace iPhoto at an OS level. Some could argue that the new interface and integration (across all of Apple's platform) offers a vastly more unified, both in look and function experience. I would happen to agree with that sentiment.

There's problems with that sentiment.

First, from effecting a smooth, professional transition, one doesn't drive the car off a cliff and hope that the replacement works - - that's the same basic mistake that Apple made with Final Cut X, so it isn't like they can claim ignorance.

Second, any sufficiently complex product will result in a variety of workflows which are of value to consumers, so not providing continuity (transition) for these is also a leadership failure. With iPhoto, the one feature that customers have screamed most loudly about was the disappearance of "stars" (variable rating system much more rich than 'flagged'): so then, has this feature been restored yet in Photos? If not, why has it not?

Third, unification is a double-edged sword, particularly in the context of iOS which must be extremely streamlined. Toolboxes contain a variety of tools for a reason, and not all workflow requirements are solved by only having a hammer.

FWIW, people appreciate change when they can readily recognize how it benefits them and their workflow needs (and/or at least are not opposed to it). Since you've claimed that Photos has enhanced your functional experience, please feel free to articulate what it changed in your workflows that was clearly beneficial.

I 100% agree with your Final Cut observation and think the 2013 Mac Pro along with FCPX.1+ is Apple attempting to reclaim some of that user base.

Marketshare was 49% back in 2007 ... and as of last year was at 20% and (rumored to be) still dropping. I've not been keeping close tabs on Adobe, but it doesn't look like they're interested in rewriting to OpenGL, particularly since Windows is a bigger market. Similarly, I've not heard much about Grand Central Dispatch for awhile now either.

EDIT: another quick thought on this, which is the motivation for improving FCX: was it really Apple's plan all along, or was this in response to getting very bad and very public negative reception? Considering how the guy who runs the "We Want a New Mac Pro" facebook page (20K+ followers) hardly gets time of day from Tim Cook, the prospects of "bad publicity risk" does seem to drive many of Apple's decisions. Deja ... AntennaGate.

I think it's an uphill battle for them but as I've just posted in a different thread I think Apple is a unique position to deliver next gen video better than others can given their advancements on mobile and further cross platform development of the Metal API.

We'll probably have to agree to disagree: I see that Apple has disregarded the upkeep of their core software ecosystem products and as such, has limited their future prospects. Some of this I personally believe is due to Ive owning more than just hardware - from what I've seen, he's a good hardware guy, but he's also ONLY a hardware guy: the software, power, and thermal management guys get ignored way too often.

 

[goMac]

Marketshare was 49% back in 2007 ... and as of last year was at 20% and (rumored to be) still dropping. I've not been keeping close tabs on Adobe, but it doesn't look like they're interested in rewriting to OpenGL, particularly since Windows is a bigger market. Similarly, I've not heard much about Grand Central Dispatch for awhile now either.

Adobe supports OpenGL already.

GCD is only one way to do threading. I don't know if Adobe supports GCD, but they're certainly supporting multicore in their applications.

Most people when they say Adobe mean Photoshop (and Illustrator.) And yes, Photoshop has poor multicore and OpenGL support. But that's because Photoshop has become an unmaintainable mess of a project. It's like a giant boat that Adobe is having trouble turning in a new direction. Competitors like Pixelmator have shown that it can be done, and done very well. Photoshop is the problem here, not the technology.

When we're talking about the Creative Suite as a whole, plenty of apps like Premiere, After Effects, Lightroom, and Audition make great use of multiple cores and OpenGL.

 

[-hh]

Adobe supports OpenGL already.

GCD is only one way to do threading. I don't know if Adobe supports GCD, but they're certainly supporting multicore in their applications.

Most people when they say Adobe mean Photoshop (and Illustrator.) And yes, Photoshop has poor multicore and OpenGL support. But that's because Photoshop has become an unmaintainable mess of a project. It's like a giant boat that Adobe is having trouble turning in a new direction. Competitors like Pixelmator have shown that it can be done, and done very well. Photoshop is the problem here, not the technology.

When we're talking about the Creative Suite as a whole, plenty of apps like Premiere, After Effects, Lightroom, and Audition make great use of multiple cores and OpenGL.

Thanks for that clarification. FWIW, my understanding was that the CUDA/OpenGL "fight" was more on the video rendering front.

BTW, now that I've finished reading up on the thread, I see that I should have emphasized one of my points a bit more strongly, namely that this discussion is not only about capability - - it is about value.

For example, I recently went through a "What If?" planning exercise for the cost of hardware replacement if one of my Mac Pro's were to have a complete meltdown. BTW, this was the machine only - - not the monitors, nor any software applications. As a reference baseline, the latter cMP (a 2012) originally costed out at ~$5K, although that number needs some adjustment, as its original drives have been updated. SWAG it to ~$5600 or so.

Well, what the exercise revealed was that a nMP with equal-or-better capabilities was possible, but would cost roughly $7600. Furthermore, to jump ship entirely and go to a Windows PC would also cost around $5600.

Now keep in mind what I said about OS's...they don't exist merely for us to admire them: they exist for us to run productivity applications in workflows. As such, the question then becomes... is OS X really worth an extra $2K per seat?

Unfortunately, $2K/seat is a pretty a tall order for "worker preferences"/etc... and the basic reason why there's this magnitude of "Apple Tax" is pretty clearly attributed to the particulars of the use case context (and the prior cMP's price tag) which happen to be a weak spot that's poorly served (price, not capability) by the nMP architecture.

-hh

 

[thefredelement]

Thank you for such a thorough reply, I really like this forum and like talking about this stuff in a well thought out manner so I appreciate what you're saying.

Re: the SATAII/III HDD vs SSD, I agree, SATA3 should have been included long ago, though to you original point regarding the 2013 MP being a success or failure, in part, based on value either gained or lost through internal storage augmentation, I don't think it should be considered because there is a viable option that is fast (and maybe faster than what was previously available) with Thunderbolt2 storage (ether SSD or with spinners and a lot of cache, or even striped).

I think we're seeing the two sides of the same coin regarding expansion and thunderbolt. PCIe is great, but there are't unlimited slots. Thunderbolt can connect a multitude of devices over it's different buses that if were internally incorporated would result in a truly monstrous enclosure. One that would also have to have many fans and accommodate cooling for the cards and the rest of the devices that shared the enclosure.

Primary subsystems of the 2013 form factor can be upgraded, such as RAM, CPU & SSD. The GPUs, while no options exist presently, would be a theoretical walk in the park replacement.

I totally understand what you're saying about Photos but both iPhoto and Aperture still run if one is so inclined to use them. (Personally the like button and albums are good enough for me). I absolutely have to admit that the first version of Photos was not good and there was a bit of a shock but I understand your sentiment, especially regarding workflow and and specifically with Aperture. While I own both iPhoto & Aperture I personally haven't used either since I've acclimated to Photos that being said I'm not a pro photog and have no idea if the last version of Aperture competes with Lightroom (or any others). Though there was plenty of notice and as mentioned, Aperture didn't just stop working. For Photos now though it seems faster to make the most common adjustments than it was in iPhoto. I am def. not the best use case for this.

The unification or the Metal API is tremendous for both iOS and OS X. Some of Apple's own "Core" series APIs have already been re-wrriten to use Metal on both platforms. Nvidia is only now starting to work with CUDA on their Tegra chips (chips that might I add always pale in comparison to Apple's offerings in real life). This is the start of a world of possibilities for both platforms and signifies Apple's direction in the GPU computational space - they haven't really adopted or participated in other standards recently or effectively. This is a big deal because it's being actively developed so close to both ARM and desktop GPUs.

Final Cut Pro X runs better on a 2013 Mac Pro than any other workstation. Apple has optimized it to specifically take advantage of the dual D series FriePros (as old as they are). Bare feats even has a benchmark showing this to be true - I think it was from last year, I've added it in a different post in this thread. I 100% agree with you that Apple's offering around FCS3/FCP7 was the beginning of a downward trend in the pro space, where I think we digress is that I think Apple took it as an opportunity to build for the future and it's taking longer to regain market share than it did to lose it. I believe Apple has seen the writing on the wall and is developing a workstation class that can accommodate many more times the cores that are currently available, knowing primarily that heat dissipation will be a huge factor with stuffing in more computational / graphics cores.

Their optimization and continued development of FCPX can be seen as the opposite of abandoning the pro space, though their next offerings will have to start moving mountains compared to the competition to start winning back pro video and I think/hope that, that is what their plan is.

True enough on the surface - - but:

First, SATA technology moved to SATA-III (6Gbps) circa 2008/09m, so not only could something faster than SATA-II been adopted in the 2010 Mac Pro update, it became downright shameful that it wasn't incorporated in the 2012. As such, we're differentiating based on false pretenses.

Second, performing incremental updates (vs wholesale system replacement) is precisely what the appeal is for things like PCIe slots in towers. Case in point: what provisions exist in the nMP to upgrade a to a higher performance subsystems? Granted, we can debate about what percent of customers do upgrades, but when there's zero provisions to upgrade at all, this too becomes a moot point.

Third, Thunderbolt certainly has speed, but it also has latency which makes it a not-quite-equal to native PCIe. Which may or may not matter for one's use case, but needs to be noted to see if it is an important differentiator or safely irrelevant.

Fourth, the performance bottleneck on traditional spinning drives is the drive itself, not the interface. As such, SATA-III is primarily only important if one wants several SSDs in drive bays so as to not give up 'too many" PCIe slots for the capability.



There's problems with that sentiment.

First, from effecting a smooth, professional transition, one doesn't drive the car off a cliff and hope that the replacement works - - that's the same basic mistake that Apple made with Final Cut X, so it isn't like they can claim ignorance.

Second, any sufficiently complex product will result in a variety of workflows which are of value to consumers, so not providing continuity (transition) for these is also a leadership failure. With iPhoto, the one feature that customers have screamed most loudly about was the disappearance of "stars" (variable rating system much more rich than 'flagged'): so then, has this feature been restored yet in Photos? If not, why has it not?

Third, unification is a double-edged sword, particularly in the context of iOS which must be extremely streamlined. Toolboxes contain a variety of tools for a reason, and not all workflow requirements are solved by only having a hammer.

FWIW, people appreciate change when they can readily recognize how it benefits them and their workflow needs (and/or at least are not opposed to it). Since you've claimed that Photos has enhanced your functional experience, please feel free to articulate what it changed in your workflows that was clearly beneficial.



Marketshare was 49% back in 2007 ... and as of last year was at 20% and (rumored to be) still dropping. I've not been keeping close tabs on Adobe, but it doesn't look like they're interested in rewriting to OpenGL, particularly since Windows is a bigger market. Similarly, I've not heard much about Grand Central Dispatch for awhile now either.



We'll probably have to agree to disagree: I see that Apple has disregarded the upkeep of their core software ecosystem products and as such, has limited their future prospects. Some of this I personally believe is due to Ive owning more than just hardware - from what I've seen, he's a good hardware guy, but he's also ONLY a hardware guy: the software, power, and thermal management guys get ignored way too often.

 

[goMac]

Thanks for that clarification. FWIW, my understanding was that the CUDA/OpenGL "fight" was more on the video rendering front.

CUDA and OpenGL are pretty much always used together in the context of what Adobe is doing. CUDA and OpenCL are at odds, but Adobe has been moving ahead with OpenCL support. My read on where they want to go in the future is they want to move past both OpenCL and CUDA to Vulkan, but Apple isn't supporting it, which Adobe has responded to by boycotting Metal.

 

[AidenShaw]

CUDA and OpenGL are pretty much always used together in the context of what Adobe is doing. CUDA and OpenCL are at odds, but Adobe has been moving ahead with OpenCL support. My read on where they want to go in the future is they want to move past both OpenCL and CUDA to Vulkan, but Apple isn't supporting it, which Adobe has responded to by boycotting Metal.

Which is a smart move, since FCPX market share is shrinking.... Why offer people options to run on a platform that your (increasingly less important) competitor only runs on?

 

[pat500000]

Wow, there is a wall of words.

So, are you saying that you needed CUDA and nMP had none so you had to buy a machine with an Nvidia card? Oh the irony...

So while the 6,1 failed to meet your needs, a different computer met them so you bought that instead, thus voting with your wallet?

And where are those nMP GPU upgrades you predicted? Pretty sure you said 12-18 months a couple years back. Are you still post dating those or finally ready to admit that they are officially "vapor-ware"?

wait..i thought you're working on that.

 

[developer13245]

A system with this architecture limitation cannot claim to be a "Pro" model:
https://forums.macrumors.com/thread...ing-for-nmp-proof-is-in-the-firmware.1876181/

Fail

 

[MacVidCards]

Some news on nMP & eGPU is coming. Should open up some eyes.

 

[hollyhillbilly]

Which is a smart move, since FCPX market share is shrinking.... Why offer people options to run on a platform that your (increasingly less important) competitor only runs on?

Please provide data that shows FCPX share shrinking since the programs launch

 

[linuxcooldude]

Please provide data that shows FCPX share shrinking since the programs launch

Good catch. Since FCPX was brand new at the time, I don't see how it could shrink below zero at that time. So on the most part we should see it go up in adoption.

 

[MacVidCards]

Talk about clinging to a semantic thread.

 

[ScottishCaptain]

The unification or the Metal API is tremendous for both iOS and OS X. Some of Apple's own "Core" series APIs have already been re-wrriten to use Metal on both platforms. Nvidia is only now starting to work with CUDA on their Tegra chips (chips that might I add always pale in comparison to Apple's offerings in real life). This is the start of a world of possibilities for both platforms and signifies Apple's direction in the GPU computational space - they haven't really adopted or participated in other standards recently or effectively. This is a big deal because it's being actively developed so close to both ARM and desktop GPUs.

Metal is already DOA.

https://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&p=188758

Note the posts from "twod", a developer working for Side Effects. Metal is simply incapable of supporting a fairly large amount of high end professional 3D applications, and this isn't going to change anytime soon because the missing features either wouldn't work or wouldn't work well on a handheld, and they can't start introducing stuff that works on OS X and not iOS after making a big deal about the API working the same on either platform.

-SC

 

[MacVidCards]

Metal is already DOA.

https://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&p=188758

Note the posts from "twod", a developer working for Side Effects. Metal is simply incapable of supporting a fairly large amount of high end professional 3D applications, and this isn't going to change anytime soon because the missing features either wouldn't work or wouldn't work well on a handheld, and they can't start introducing stuff that works on OS X and not iOS after making a big deal about the API working the same on either platform.

-SC

OH COME ON ! It does great Emojis on iPad Semi-Pro and Angry Birds has never looked angrier ! You people.

Apple moving the goalposts every 6 months keeps the devs on their toes.

 

[thefredelement]

Metal is already DOA.

https://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&p=188758

Note the posts from "twod", a developer working for Side Effects. Metal is simply incapable of supporting a fairly large amount of high end professional 3D applications, and this isn't going to change anytime soon because the missing features either wouldn't work or wouldn't work well on a handheld, and they can't start introducing stuff that works on OS X and not iOS after making a big deal about the API working the same on either platform.

-SC

Metal has been released on OS X for less than 6 months, to call it DOA is greatly premature. There are also posts in the thread you referenced mentioning how benchmarks discussed don't even use current Apple hardware. I'm not going to scour the internet to link to other forums but instead you can read more about Metal here:

https://developer.apple.com/library...nce/index.html#//apple_ref/doc/uid/TP40015356

And also note, Metal is the future on the Apple platform, maybe one day until it isn't, but that day certainly isn't today.

 

[MacVidCards]

Metal has been released on OS X for less than 6 months, to call it DOA is greatly premature. There are also posts in the thread you referenced mentioning how benchmarks discussed don't even use current Apple hardware. I'm not going to scour the internet to link to other forums but instead you can read more about Metal here:

https://developer.apple.com/library...nce/index.html#//apple_ref/doc/uid/TP40015356

And also note, Metal is the future on the Apple platform, maybe one day until it isn't, but that day certainly isn't today.

Maybe just link us directly to the Apple PR/BS department.

You do realize they were positively RAVING about OpenCl 2 years ago, yes? Now it's day old sushi they can't distance themselves from quickly enough.

Time for another round of Kool Aid ! Doubles!