computer (laptop) design is pretty much over, it's been mastered ridiculously. It's freaking like a front room tv, you get it, it works, you go...people's speed requirements, are like nickel and dime-ing a couple benchmark scores here and there, we used to have bench scores for laptops in the 500s, we are currently in the 12,500!s now days with a laptop and it's leveled off, I mean come on, what is it you CAN'T DO?
p.s. oh I forgot, Crysis, or no, now it's Oculus....
I cannot emphasize this enough. Intel integrated graphics stink. Try using Duet on a Mac laptop with the 12.9" iPad Pro. It is an atrocious experience for what would otherwise be a fantastic tool. The lag simply isn't there when the iPad Pro is attached to an iMac with dedicated GPU. And the desktop doesn't heat up like the laptop when using Duet.
You do know MacBook Pro is for professional market, at least this is Apple's intension. If you try to use any Adobe program on iPad Pro without optional keyboard & mouse, you will want to kill yourselves because you can't use any shortcut keys.
In addition, almost all professional program is designed for x86 base system.
It will be too much work and less return for them to rewrite/plot to ARM base processor. At least for the next fees years
There's no reason why Apple can't offer ALL their Pro machines with discrete GPU's and faster/better overall specs. Apple customers will pay higher prices and provide the margins Apple needs to continue their investment in the Mac platform.
iPhones, iPads and iOS devices are great and all that, but Apple needs to get on the ball, and get its' act together when it comes to the Mac line.
If phones and tablets are cars, then Macs are trucks, like Steve Jobs said many years ago, and certain people do want and do still need powerful trucks.
Apple makes most of its profit from iPhones today, but that is no excuse to neglect their computer and Mac line.
Mac Pro won't be upgraded in GPGPUs until Vega arrives. AMD has already stamped out Vega with HBM2 memory up to 32GB 2TB bandwidth GPGPUs fro the 1H2017.
Apple will get them early as a custom ASIC design.
The 14nm Radeon Polaris RX 8GB in a Macbook Pro would be a beast in a laptop of any brand.
AMD has yet to release any Mobile version of the RX, but Apple will most likely be the first this September.
Thanks for the well rounded and informative article.
My thoughts are that Apple's reliance on Intel caused a sever delay in updating. Then the Skylake chipset's problems after release didn't help. After that, with so many new technologies able to be implemented it looked better to delay rather than give out another half-assed update.
It's clear I think that the next PC standard will not be defined by Apple. But I think most people will be perfectly happy using a $30 ARM 64 bit super computer running Linux Ubuntu.
But I think most people will be perfectly happy using a $30 ARM 64 bit super computer running Linux
There, fixed it for you.
Using Linux means you have a choice. I prefer CentOS + Cinammon.
I'd go with an ARM Cpu provided :- that Adobe CS was available and there was at least 2TB of storage in the device. Cloud storage is useless for my use case.
It's clear I think that the next PC standard will not be defined by Apple. But I think most people will be perfectly happy using a $30 ARM 64 bit super computer running Linux Ubuntu.
But I think most people will be perfectly happy using a $30 ARM 64 bit super computer running Linux
There, fixed it for you.
Using Linux means you have a choice. I prefer CentOS + Cinammon.
I'd go with an ARM Cpu provided :- that Adobe CS was available and there was at least 2TB of storage in the device. Cloud storage is useless for my use case.
It's clear I think that the next PC standard will not be defined by Apple. But I think most people will be perfectly happy using a $30 ARM 64 bit super computer running Linux Ubuntu.
But I think most people will be perfectly happy using a $30 ARM 64 bit super computer running Linux
There, fixed it for you.
Using Linux means you have a choice. I prefer CentOS + Cinammon.
I'd go with an ARM Cpu provided :- that Adobe CS was available and there was at least 2TB of storage in the device. Cloud storage is useless for my use case.
On the positive side, I've got a 2013 rMBP with a dedicated GPU and I'm not in the least tempted to update. In the old days, I'd worry about what was coming out 6 months after purchase.
Know what you man analogjack. The old man was looking at how old his 2013 MBP I s getting. I told him it is the same as the current Mac with marginal performance difference. No need to upgrade.
wood12-8 said
"Apple can help make Mackbook Pro thinner/lighter(for college kids carrying heavy backpack with tons of textbooks along with macbook pro,etc)".
Most College books are available as pdf's now which are not that heavy
Apple is missing an opportunity to produce "Pro" machines. There is no reason they can't make both consumer and pro machines, other than they don't want to be bothered. Just license OS X, it runs faster and better on other peoples hardware. I have been using Macs since 1987... and I am now forced to use Windoze because of performance, it is embarrassing. Not one computer with desktop class graphics.. and no i am not counting my 2013 Mac Pro, my 2013 hackintosh runs much faster and has upgradable graphics.
I'm not willing to bother with the hackintosh. I did that initially when the Mac Pro still had a PCIe model just to see if it was viable.
While it's probably less irritating to configure than Linux, the legality of it makes it an awful choice because it would be requiring running potentially dangerous copies of the macOS operating system obtained from unsanctioned sources. Even then, hardware support is hard to predict.
Apple really does need to do a "**** or get off the pot" when it comes to the "Pro" hardware. If they're not willing to put out serious Pro hardware, they need to either say so and axe the "pro" lines, or they need to put out some high-end upgradable hardware that gamers and professionals would envy, not feel embarrassed about.
Apple is missing an opportunity to produce "Pro" machines. There is no reason they can't make both consumer and pro machines, other than they don't want to be bothered. Just license OS X, it runs faster and better on other peoples hardware. I have been using Macs since 1987... and I am now forced to use Windoze because of performance, it is embarrassing. Not one computer with desktop class graphics.. and no i am not counting my 2013 Mac Pro, my 2013 hackintosh runs much faster and has upgradable graphics.
Producing high-end computers isn't much of an opportunity these days. Higher end machines years ago sold in a higher proportion of overall sales because computers were so slow and lots of people needed a minimum performance threshold that was higher up. As computers got faster and cheaper, the performance threshold hit a lower entry point. Nowadays the dominant computers are ~$500 PC laptops.
The amount of people involved in high-end production is very small. The number of blockbuster movies made every year is in the hundreds (low thousands worldwide):
"In 2004, roughly 490 films were released on fewer than 1,000 screens, according to data compiled by the National Association of Theatre Owners (NATO). Last year, that number ballooned to 563 movies."
The number of editors working those movies is similarly in the hundreds or low thousands. Apple sells ~25 million Macs every year and the vast majority are lower-end laptops.
This shouldn't come as a surprise, the reason that people who work in high-end computing jobs only make a high salary because there are so few of them. If there were loads of them, their salaries would be much lower due to increased competition.
When you look at high-end desktop companies like Boxx, you can see their client list:
"Categorized under Computers Manufacturers. Our records show it was established in 1995 and incorporated in Texas. Current estimates show this company has an annual revenue of $1 to 2.5 million and employs a staff of approximately 1 to 4."
Even if you took their revenue to be $20m and each product was $2k, that's 10,000 units. Most of the hardware the big movie companies buy is server hardware.
The consumer-level high-end desktop/laptop market (enthusiast market) is geared towards gamers because there are as many PC gamers as console gamers. This involves cramming desktop-class GPUs into small spaces by underclocking them and attaching all manner of crazy cooling solutions. If you tried to run a laptop with a 100W GPU at full load on battery power, it wouldn't last an hour and it will sound like a hair-dryer all the time.
Apple builds machines to best suit their use case. Laptops are portable so they need good battery life, they need to be fairly quiet in operation and not get too hot. They tend to use <40W GPUs. This is fine for a laptop and the upcoming 35W AMD M480 performs well enough:
That sits somewhere between an NVidia 960M and 970M and is roughly double the R9 M370X. The Skylake GT4e doesn't look like it will match the M370X but it's good for OpenCL. The 480M is a fair bit short of the desktop-class 'notebook' GPUs:
The performance there is about 4x the 480M but it's 180W TDP. This is where the iMac comes in. Apple uses a GPU like that in the iMac because it can support a higher power GPU and it tends to be around 3x faster than the MBP.
The Mac Pro is the next step up and gets dual high-end full desktop GPUs and should be 2-3x the iMac. The amount of people who need to buy these is tiny as mentioned earlier. It's in the tens of thousands per quarter. They are high margin but very low volume. Even with another refresh, it will be possible to build enthusiast hardware cheaper and with faster GPUs. 1080 SLI for $1200 will likely outperform the next dual AMD Mac Pro option at $1000 so a $2k gaming/enthusiast rig could outperform at $4k Mac Pro. The gaming/enthusiast rig won't have anywhere near the same margins as the Mac Pro so there's little point in Apple pursuing it.
Apple might sell more of the Mac Pro if they focused on the enthusiast market rather than the workstation audience. They could build a machine with a 6-core CPU and dual NVidia 1080 for $3.5k-4k and still get decent margins but the risk there is whether the enthusiast market would pay those prices. The workstation market would. If they dropped one of the GPUs then they could hit below $3k. You have to assume that a company that manages to sell 25 million units every year and makes more profit margin than any other company knows a bit about the market they are selling to.
Over the next few years, people are going to keep shifting to lower and lower machines the faster they get. The next iMac should outperform a desktop 980 so if all that people wanted was an enthusiast desktop in 2014 with a quad-i7 and 980 GPU for ~$2k, now they can get one in 2016 and it will bundle a Retina display.
The Mac Pro can get up to 18-core CPU and dual Fiji workstation GPUs that perform about 14-16TFLOPs. The newer architectures offer better performance-per-watt but the highest-end parts tend to be a generation behind. The Xeons will still be Broadwell.
The Mac Pro only exists to meet performance demands, mostly for graphics power. Tim Sweeney who has been at the Apple keynote and is CEO of Epic who makes Unreal game engine estimates 40TFLOPs will achieve photoreal graphics in real-time:
This is roughly quad 1080 SLI. With improvements to software and fixed-function hardware, computers can reach this in the next 4 years or so and there's not much more to do. Once you can achieve photoreal VR (4K, 90FPS), there's not a level above this because you can't get more real and it will never be necessary to do all the computation in real-time.
It's the buyers who are going to shape the computer market, not the sellers. NVidia and AMD will be around as long as they can make a profit. Look at AMD after the launch of Polaris:
but look at the revenues. $1.4b in a quarter. If you divide that by $200 GPUs, you get 7 million units across all PCs. If you divide it by $600 GPUs (flagship level), it's 2.3m units. Hardly anyone is buying enthusiast level GPUs, relative to the overall computer industry. NVidia's worldwide enthusiast GPU sales are lower than Mac sales. NVidia made $5b revenue last year, that's equivalent to 8 million 980ti GPUs at most. This makes it less than 5% of the overall market, mid-range GPUs expand it out a bit.
If Apple was to target this and even manage to capture the entire market, it's a few million units. Multiply that by $500 net margins to get $2.5b. They made over $50b last year and they wouldn't get anywhere near the entire enthusiast market, they'd get 20% at most ($500m). If it was a viable opportunity for Apple, they'd go after it. Instead the MBP and iMac target the much larger mid-range segment and that works pretty well for them so that's what to expect going forward.
Apple is missing an opportunity to produce "Pro" machines. There is no reason they can't make both consumer and pro machines, other than they don't want to be bothered. Just license OS X, it runs faster and better on other peoples hardware. I have been using Macs since 1987... and I am now forced to use Windoze because of performance, it is embarrassing. Not one computer with desktop class graphics.. and no i am not counting my 2013 Mac Pro, my 2013 hackintosh runs much faster and has upgradable graphics.
Producing high-end computers isn’t much of an opportunity these days.
So there’s just not much money to be made in high-end machines. I get it. Why should Apple bother with such a small market? And while we’re at it why should Apple bother with the low-end smartphone market when there’s little profit to be made? The problem here with AI commenters claiming otherwise is their demographics in that they think their needs and wants represent the majority market when clearly the truth lies elsewhere.
Comments
If phones and tablets are cars, then Macs are trucks, like Steve Jobs said many years ago, and certain people do want and do still need powerful trucks.
Apple makes most of its profit from iPhones today, but that is no excuse to neglect their computer and Mac line.
My thoughts are that Apple's reliance on Intel caused a sever delay in updating. Then the Skylake chipset's problems after release didn't help. After that, with so many new technologies able to be implemented it looked better to delay rather than give out another half-assed update.
The Mac sales are 20% of the iPhone sales. Not an insignificant amount that can be ignored or cast aside. In 2015 Apple(for Mac only) was in the top 5 PC OEMs for the first time in nearly 20 years(https://en.wikipedia.org/wiki/Market_share_of_personal_computer_vendors).
For general use the Mac is still a great machine. The only worry is for graphic intense and outlier professional use.
There, fixed it for you.
Using Linux means you have a choice. I prefer CentOS + Cinammon.
I'd go with an ARM Cpu provided :- that Adobe CS was available and there was at least 2TB of storage in the device. Cloud storage is useless for my use case.
There, fixed it for you.
Using Linux means you have a choice. I prefer CentOS + Cinammon.
I'd go with an ARM Cpu provided :- that Adobe CS was available and there was at least 2TB of storage in the device. Cloud storage is useless for my use case.
There, fixed it for you.
Using Linux means you have a choice. I prefer CentOS + Cinammon.
I'd go with an ARM Cpu provided :- that Adobe CS was available and there was at least 2TB of storage in the device. Cloud storage is useless for my use case.
While it's probably less irritating to configure than Linux, the legality of it makes it an awful choice because it would be requiring running potentially dangerous copies of the macOS operating system obtained from unsanctioned sources. Even then, hardware support is hard to predict.
Apple really does need to do a "**** or get off the pot" when it comes to the "Pro" hardware. If they're not willing to put out serious Pro hardware, they need to either say so and axe the "pro" lines, or they need to put out some high-end upgradable hardware that gamers and professionals would envy, not feel embarrassed about.
The amount of people involved in high-end production is very small. The number of blockbuster movies made every year is in the hundreds (low thousands worldwide):
http://variety.com/2015/film/news/hollywood-making-too-many-movies-1201526094/
"In 2004, roughly 490 films were released on fewer than 1,000 screens, according to data compiled by the National Association of Theatre Owners (NATO). Last year, that number ballooned to 563 movies."
The number of editors working those movies is similarly in the hundreds or low thousands. Apple sells ~25 million Macs every year and the vast majority are lower-end laptops.
This shouldn't come as a surprise, the reason that people who work in high-end computing jobs only make a high salary because there are so few of them. If there were loads of them, their salaries would be much lower due to increased competition.
When you look at high-end desktop companies like Boxx, you can see their client list:
http://www.boxx.com/about-us/our-customers
This company was bought by a private equity firm:
http://www.bizjournals.com/austin/blog/techflash/2016/01/boxx-technologies-bought-by-dallas-firm.html
http://www.craftsmancapitalpartners.com/portfolio.html
http://www.craftsmancapitalpartners.com/investment-criteria.html
Their investment criteria states the companies they invest in should have earnings below $5m.
http://www.manta.com/c/mmck1t6/boxx-technologies
"Categorized under Computers Manufacturers. Our records show it was established in 1995 and incorporated in Texas. Current estimates show this company has an annual revenue of $1 to 2.5 million and employs a staff of approximately 1 to 4."
Even if you took their revenue to be $20m and each product was $2k, that's 10,000 units. Most of the hardware the big movie companies buy is server hardware.
The consumer-level high-end desktop/laptop market (enthusiast market) is geared towards gamers because there are as many PC gamers as console gamers. This involves cramming desktop-class GPUs into small spaces by underclocking them and attaching all manner of crazy cooling solutions. If you tried to run a laptop with a 100W GPU at full load on battery power, it wouldn't last an hour and it will sound like a hair-dryer all the time.
Apple builds machines to best suit their use case. Laptops are portable so they need good battery life, they need to be fairly quiet in operation and not get too hot. They tend to use <40W GPUs. This is fine for a laptop and the upcoming 35W AMD M480 performs well enough:
http://www.nextpowerup.com/news/28617/amd-radeon-rx-480m-gets-detailed/
That sits somewhere between an NVidia 960M and 970M and is roughly double the R9 M370X. The Skylake GT4e doesn't look like it will match the M370X but it's good for OpenCL. The 480M is a fair bit short of the desktop-class 'notebook' GPUs:
http://www.notebookcheck.net/NVIDIA-GeForce-GTX-1080-Laptop.171212.0.html
The performance there is about 4x the 480M but it's 180W TDP. This is where the iMac comes in. Apple uses a GPU like that in the iMac because it can support a higher power GPU and it tends to be around 3x faster than the MBP.
The Mac Pro is the next step up and gets dual high-end full desktop GPUs and should be 2-3x the iMac. The amount of people who need to buy these is tiny as mentioned earlier. It's in the tens of thousands per quarter. They are high margin but very low volume. Even with another refresh, it will be possible to build enthusiast hardware cheaper and with faster GPUs. 1080 SLI for $1200 will likely outperform the next dual AMD Mac Pro option at $1000 so a $2k gaming/enthusiast rig could outperform at $4k Mac Pro. The gaming/enthusiast rig won't have anywhere near the same margins as the Mac Pro so there's little point in Apple pursuing it.
Apple might sell more of the Mac Pro if they focused on the enthusiast market rather than the workstation audience. They could build a machine with a 6-core CPU and dual NVidia 1080 for $3.5k-4k and still get decent margins but the risk there is whether the enthusiast market would pay those prices. The workstation market would. If they dropped one of the GPUs then they could hit below $3k. You have to assume that a company that manages to sell 25 million units every year and makes more profit margin than any other company knows a bit about the market they are selling to.
Over the next few years, people are going to keep shifting to lower and lower machines the faster they get. The next iMac should outperform a desktop 980 so if all that people wanted was an enthusiast desktop in 2014 with a quad-i7 and 980 GPU for ~$2k, now they can get one in 2016 and it will bundle a Retina display.
The Mac Pro can get up to 18-core CPU and dual Fiji workstation GPUs that perform about 14-16TFLOPs. The newer architectures offer better performance-per-watt but the highest-end parts tend to be a generation behind. The Xeons will still be Broadwell.
The Mac Pro only exists to meet performance demands, mostly for graphics power. Tim Sweeney who has been at the Apple keynote and is CEO of Epic who makes Unreal game engine estimates 40TFLOPs will achieve photoreal graphics in real-time:
http://www.gamespot.com/articles/tim-sweeney-criticizes-microsoftoculus-and-talks-p/1100-6441734/
This is roughly quad 1080 SLI. With improvements to software and fixed-function hardware, computers can reach this in the next 4 years or so and there's not much more to do. Once you can achieve photoreal VR (4K, 90FPS), there's not a level above this because you can't get more real and it will never be necessary to do all the computation in real-time.
It's the buyers who are going to shape the computer market, not the sellers. NVidia and AMD will be around as long as they can make a profit. Look at AMD after the launch of Polaris:
http://www.amd.com/en-us/press-releases/Pages/press-release-2016jul21.aspx
$69m net income when they are $2b in the hole with debt, non-GAAP was a net loss of $40m. NVidia is doing better:
http://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-second-quarter-fiscal-2017
but look at the revenues. $1.4b in a quarter. If you divide that by $200 GPUs, you get 7 million units across all PCs. If you divide it by $600 GPUs (flagship level), it's 2.3m units. Hardly anyone is buying enthusiast level GPUs, relative to the overall computer industry. NVidia's worldwide enthusiast GPU sales are lower than Mac sales. NVidia made $5b revenue last year, that's equivalent to 8 million 980ti GPUs at most. This makes it less than 5% of the overall market, mid-range GPUs expand it out a bit.
If Apple was to target this and even manage to capture the entire market, it's a few million units. Multiply that by $500 net margins to get $2.5b. They made over $50b last year and they wouldn't get anywhere near the entire enthusiast market, they'd get 20% at most ($500m). If it was a viable opportunity for Apple, they'd go after it. Instead the MBP and iMac target the much larger mid-range segment and that works pretty well for them so that's what to expect going forward.