AppleInsider podcast talks Mac Pro, the future of PowerVR, Clips & more

"'Constant Negativity' From Pro Users Led Apple to Develop Modular Mac Pro, Which May Not Ship Until 2019" https://www.macrumors.com/2017/04/06/mac-pro-may-not-ship-until-2019/ Get this straight: criticising Apple when doing WRONG is not bad, as some believe around here believe (probably because they have Apple shares)!

I hope Apple doesn't just rush out a new MacPro because people are complaining about the current version. Take the time to get it right and they won't be in the position again. I'm pretty confident they will do just that...its not really like Apple to just rush something out the door. 

Good podcast,  as usual! :) Mikey does well when he's the host & DED's depth of knowledge and insights to what Apple may be considering based on past performance is always interesting.

DED: "...people basically creating a 'dossier' of themselves and just giving it to advertisers." Facebook in one sentence. Brilliant.

Best 

appex said:

"'Constant Negativity' From Pro Users Led Apple to Develop Modular Mac Pro, Which May Not Ship Until 2019" https://www.macrumors.com/2017/04/06/mac-pro-may-not-ship-until-2019/ Get this straight: criticising Apple when doing WRONG is not bad, as some believe around here believe (probably because they have Apple shares)!

It's dishonest and disingenuous to throw out false accusations without backing them up.

The most recent, opinions made about Mac Pro were not exactly white-washing, free of criticism: http://appleinsider.com/articles/17/03/05/editorial-the-future-of-apples-macintosh

Mac Pro

The Late 2013 release of Mac Pro may have been a mistake. Its design wasn't readily upgradable, but Apple also lacked the sales volumes to warrant regular significant update cycles. If a cycle is too long, the benefits of product cycles described above begin to evaporate. It may have been better for Apple to have designed a system other vendors could upgrade, with room for standard PCIe graphics cards and perhaps even CPU packages.

It appears that Apple mistakenly approached the workstation class PC product segment with the same integrated design skills that worked so well for iMac and iPhone. This approach previously failed in designing Xserve, which while expandable didn't address the basics that rack mounted server buyers find important while focusing on features that market doesn't care about (including an easy to use, simple user interface).

There are some potential ways Apple could rethink its Mac Pro strategy. The simplest change would be to open up its existing design to accommodate third party CPU and GPU processor upgrades. One path to doing this is enabled by Thunderbolt 3, which Apple demonstrated an early affinity for last fall in its design of new MacBook Pros.

By enabling third party PCIe enclosures connected with Thunderbolt 3, Apple could address the needs of high-end users of both MacBook Pros and desktop Mac Pro models. The reason why this hasn't happened yet is obvious: Thunderbolt 3 wasn't previously available. Simply updating the existing Mac Pro with Thunderbolt 3 would offer networked benefits for all Pro Mac users, desktop or mobile.

However, designing an updated new Mac Pro doesn't change the fact that there is currently very limited demand for Mac desktops. Part of this is related to Apple's constrained strategy for its macOS platform, which focuses entirely on the Mac hardware Apple builds itself."

macxpress said:

I hope Apple doesn't just rush out a new MacPro because people are complaining about the current version. Take the time to get it right and they won't be in the position again. I'm pretty confident they will do just that...its not really like Apple to just rush something out the door. 

There isn't an easy way to do that though because different tasks have different needs. Musicians don't need GPUs, they need quiet machines that have good IO performance and multi-threading. Designers need lots of RAM, fast storage, a decent GPU but not so much CPU. Video editors need fast CPUs for encoding, fast storage and lots of storage capacity, sometimes a decent GPU. Video compositors need multiple GPUs and CPUs for encoding. 3D needs as many CPUs as possible and a fast GPU for real-time. You can't design a single system that suits them all perfectly well; if you adequately cover the needs of all of them then they all end up with a box that has compromises against their needs. This is why manufacturers like HP, Boxx etc have different boxes:

http://www8.hp.com/us/en/workstations/z2mini.html
http://www.boxx.com/products/workstations
http://www.boxx.com/products/mobile-workstations

They don't sell one box for every use case but one-size fits all is part of Apple's minimalist design. One way they might be able to go is to have a box with 4 slots that can take either CPUs or GPUs, like this but it would have to be a bit bigger:

https://www.akitio.com/desktop-storage/akitio-thunder2-quad


Every system needs 1 CPU and 1 GPU minimum but something like that would allow putting 1 CPU board plus 1 GPU board plus 2 IO boards for musicians, extra storage capacity for video editing and IO; 1 CPU plus 3 GPUs for compositing; 3 CPUs plus 1 GPU for 3D.

The biggest complaint about the cylinder model was being locked into having an AMD GPU for the life of the machine with no options, most people would rather have NVidia GPUs. This can be solved with a similar cylinder design by just putting GPU slots where the custom GPUs are and having it be a bit taller. For people who need storage, they take one of the GPUs out and have storage bays there. Single CPUs will be going to 24-core/48-thread so multiple CPUs aren't essential.

Now, while these kind of setups would be more flexible, the other part of the problem is the price. Even though people always say that money is no object in this market, money is a factor for everyone and most workstations sold aren't above $2k, Apple's start at $3k. The following site estimated the ASP of Intel's server/workstation chips and it came to about $550:

https://www.fool.com/investing/general/2016/02/10/computing-the-average-selling-price-of-an-intel-co.aspx

The chips in the 12-core MP alone cost nearly $3k. The potential unit sales are limited by what the target market will pay for the product and margins impact this. Apple can build amazingly powerful machines (quad 24-CPU, 8 Titan GPU) and sell them for $20k and everybody will go 'wow, that's more like it' and then nobody buys it anyway because it costs too much so there's no point in doing that.

Their potential sales are about $1b, which if the unit volume drops 50% for every $1k increase in price means something like 150k x $3k, 75k x $4k, 35k x $5k, 17k x $6k, 8k x $7k. That would be just under 3x unit volume where they are now. They can't go to 10x to match HP/Dell units without offering workstations in the $1000-1500 range. When you get high enough in price, the number of units would just tail off to insignificant numbers.

When Steve Jobs handled the XServe problem, he just told it like it is - they cancelled it because 'nobody was buying them'. It wasn't that Apple made the server wrongly or badly, the price plus the software plus the refresh cycle just didn't match the market. The Mac Pro price doesn't match the lower performance workstation market and the Mac Pro performance and refresh cycle doesn't match with the higher performance workstation market. Although people could argue that is doing it badly/wrongly, it's like how the music market wants everything for free. There's a point where it's not sustainable and people just exit the market. HP wanted to abandon the PC market for this reason.

Apple's solution appears to be to split into two models, which is bring the iMac Pro for the lower performance, higher unit volume workstation market and bring some standard box for the higher performance market. If it's a box, it might not even be sold mainstream, it could be a build to order service because the units are so low. The iMac Pro wouldn't satisfy the people who want the headless form factor though and a large box wouldn't be as suitable for designers/photographers/musicians/editors.

The problem is when people aren't happy, everybody says Apple isn't doing it right but they mean different things. The way Apple stops the complaints these days is to tell them they're all correct and give empty promises that something far in the future will fix it. Steve Jobs just spelled it out and what's crazy is that people still think these issues are a result of modern Apple when Steve Jobs was going to abandon the pro line years ago:

http://kensegall.com/2013/08/apples-evolving-view-of-pro/

"Could it possibly be? Would Apple ever even think about saying goodbye to the pro market?

I hope you’re sitting down for this, but Steve Jobs did in fact once consider that very option.

This was back in the days when iMac had established itself as a global bestseller. During one of the agency’s regular meetings with Steve, he shared that he was considering killing the pro products.

His rationale was as you might expect: consumer products have an unlimited upside, while pro products are aimed at a niche market that eats up major resources.

Obviously, the pro market has value for Apple, even if its numbers are relatively small. Pros are opinion leaders, influencers and evangelists. Their love of Apple shows up in the purchase decisions of friends, family and colleagues.

So Steve ultimately renewed his commitment to the pros—but he never said that this commitment wouldn’t evolve. Clearly Apple has changed its thinking about the pro market, and how it can best serve its pro users."

There's nothing to be gained for Apple to say to customers that they are no longer a sustainable part of their business model. They aren't looking to argue with or alienate their customers and fans so all they can do is be polite and tell them what they want to hear. Somewhere along the line they will deliver products that people can be happy with but buyers need to understand that businesses can't always revolve around individual needs when they are building mass market products and building it yourself is actually the only way you can get what you want.