Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 7,1 Why Apple is advertising Mac Pro 2019 with Photoshop?
- Thread starter mavericks7913
- Start date
- Sort by reaction score
Yes, but the other parties sell both "thin and light" laptop models and "portable workstation" laptop models. Apple only offers anemic laptops.
My old HP was 3 times as thick as my MacBook Pro and you could toast marshmallows with that thing. I wonder what you would get if someone designed their version of the ultimate MacBook Pro.
Cooling can always be better. The no go zone for me is when it exceeds the ability to go as carry on on an aeroplane and when it can't fit into any kind of camera bag.
But did it throttle?
I have two work laptops.
One is a thin-and-light Lenovo T480s touch-screen. Quad core, 24 GiB RAM, 1 TiB NVMe SSD, 14" full HD(1080p) screen, 4G LTE modem. It's my everyday laptop.
My secondary laptop is a P52. Hex core Xeon, 64 GiB ECC RAM, 1 TiB NVMe SSD, 15.6" 4K screen, 4G LTE modem. Mainly used when I need to do an on-site customer demo.
It doesn't fit into my everyday day pack - it has to go in the roller bag. No big deal, but I always need to have the T480s in my day pack, and pack the P52 in the roller bag if needed.
Last edited:
Nice. I still plan on having a windows setup but for a different workflow than my Mac workflow.
Mine isn't a workflow issue (or maybe it is). I sometimes need the P52 and the 64 GiB of RAM for presentations. For everyday use I love the lighter touch-screen T480s.
**EDITED as I referred to the wrong poster and added more technical info**
Hi mavericks7913,
The Xeon CPU that Apple tested has yet to be released and is a generation ahead of currently available CPUs. It's boost clock is 4.4GHz, and as it is a newer generation - the Instructions Per Cycle/Clock (IPC) may be more efficient than those CPUs previously tested by Puget Systems. In other words, the New Xeon W 4.4GHz max boost could be more efficient "per clock"* than current generation 4.3GHz boost clock (iMac Pro Xeon 14 Core model has 33.25MB cache [Puget System's test], the 18 core model has 42.75MB cache [Apple's Test] - both iMac Pro models boost at 4.3GHz), and maybe even the i9 9900K (5.0GHz boost) used in Puget System's Testing quoted in Post #1.
This may be verified by Puget System's testing in the future.
*Even if the IPC were the same between the current Xeons used in the tests and new model Xeons - there is still a 0.1GHz difference favouring the upcoming generation of Xeon CPUs in the New Mac Pro, as well as higher cache (66.5MB).
The to-be-released Mac Pro also has a 300W rated heatsink - "probably" more superior
than offered by the iMac Pro models - this would remove throttling at boosted frequencies from the CPU which could occur in the iMac Pro - which may also favour the New Mac Pro.
This may also be confirmed by using a sustained CPU heavy benchmark on a few Mac and PC models.
Considering that both the iMac and to-be released Mac Pro models tested has a minimum of 256GB RAM, so it could be assumed that very large files were used (a heavy work load), and that the extra RAM may have tipped things in favour of the new Mac Pro.
Apple compared the to-be-released Mac Pro with previous/current generation models on their CPU model (Post #1)
Most customers tend to identify their model by their CPU specs.
For example - People are more likely to say that they have "the 12 Core Trash Can" model rather than saying "I have the (dual) D700 trash can model^" - even if they have both the 12 CPU Cores and dual D700 GPUs in their trash can Mac Pro.
^Other trash can models with fewer CPU cores could be configured with the dual D700 graphics.
How customers identify their computer model is usually a question of context.
As a result, Apple quoted the CPU spec of each of the model tested in the graphic you provided in Post #1 instead of the GPU specs in the models used in their tests.
In theory, Apple could have stated in the graphic (attached in Post #1) that they used the following 3 models:
New Mac Pro - Dual AMD Radeon Pro Vega II graphics with Infinity Fabric Link and 32GB of HBM2 each (128GB total).
iMac Pro - Radeon Pro Vega 64X graphics with 16GB of HBM2.
Current Mac Pro - Dual AMD D700 Graphics with 6GB of vRAM/GDDR5 RAM each (12GB total).
However, as mentioned previously, most of Apple's customers would more readily identify with the CPU model used in each test than the GPU used.
It is probably correct that Photoshop would not used more than 8 cores, but the new Mac Pro had a much beefier GPU set up and more RAM (384GB compared to Puget System's 128GB RAM).
In William Payne's testing, the Vega Pro GPU in your MBP had a higher % utilisation (90%) compared to your CPU utilisation (20-50% - although clocked at 4.0GHz).
As has been previously stated in above posts - the actual testing format used by Apple has not been released, although anyone can download the Puget System's test - so it is hard to create an apples-to-apple's comparison as 
's testing may have been more GPU focused (taking advantage of the dual Vega Pros and 128GB HBM2 - 8x more HBM2 than in the iMac Pro tested [16GB HBM2]).
It would be better to wait until the new Mac Pro is released and then run Puget System's benchmark first; and then maybe use a sustained heavier CPU based test to see if the differences in heatsinks between the iMac Pros and new Mac Pro (as well as a few PCs) have an influence in sustained boost CPU clocks.
Dominic
===================================================================
My background - I am a Pharmaceutical Scientist as well as being a computer hobbyiest (in my early 30s) - SoyCapitanSoyCapitan may have more IT experience than I have been alive LOL.
Hi mavericks7913,
The Xeon CPU that Apple tested has yet to be released and is a generation ahead of currently available CPUs. It's boost clock is 4.4GHz, and as it is a newer generation - the Instructions Per Cycle/Clock (IPC) may be more efficient than those CPUs previously tested by Puget Systems. In other words, the New Xeon W 4.4GHz max boost could be more efficient "per clock"* than current generation 4.3GHz boost clock (iMac Pro Xeon 14 Core model has 33.25MB cache [Puget System's test], the 18 core model has 42.75MB cache [Apple's Test] - both iMac Pro models boost at 4.3GHz), and maybe even the i9 9900K (5.0GHz boost) used in Puget System's Testing quoted in Post #1.
This may be verified by Puget System's testing in the future.
*Even if the IPC were the same between the current Xeons used in the tests and new model Xeons - there is still a 0.1GHz difference favouring the upcoming generation of Xeon CPUs in the New Mac Pro, as well as higher cache (66.5MB).
The to-be-released Mac Pro also has a 300W rated heatsink - "probably" more superior
than offered by the iMac Pro models - this would remove throttling at boosted frequencies from the CPU which could occur in the iMac Pro - which may also favour the New Mac Pro.This may also be confirmed by using a sustained CPU heavy benchmark on a few Mac and PC models.
Considering that both the iMac and to-be released Mac Pro models tested has a minimum of 256GB RAM, so it could be assumed that very large files were used (a heavy work load), and that the extra RAM may have tipped things in favour of the new Mac Pro.
Apple compared the to-be-released Mac Pro with previous/current generation models on their CPU model (Post #1)
Most customers tend to identify their model by their CPU specs.
For example - People are more likely to say that they have "the 12 Core Trash Can" model rather than saying "I have the (dual) D700 trash can model^" - even if they have both the 12 CPU Cores and dual D700 GPUs in their trash can Mac Pro.
^Other trash can models with fewer CPU cores could be configured with the dual D700 graphics.
How customers identify their computer model is usually a question of context.
As a result, Apple quoted the CPU spec of each of the model tested in the graphic you provided in Post #1 instead of the GPU specs in the models used in their tests.
In theory, Apple could have stated in the graphic (attached in Post #1) that they used the following 3 models:
New Mac Pro - Dual AMD Radeon Pro Vega II graphics with Infinity Fabric Link and 32GB of HBM2 each (128GB total).
iMac Pro - Radeon Pro Vega 64X graphics with 16GB of HBM2.
Current Mac Pro - Dual AMD D700 Graphics with 6GB of vRAM/GDDR5 RAM each (12GB total).
However, as mentioned previously, most of Apple's customers would more readily identify with the CPU model used in each test than the GPU used.
It is probably correct that Photoshop would not used more than 8 cores, but the new Mac Pro had a much beefier GPU set up and more RAM (384GB compared to Puget System's 128GB RAM).
In William Payne's testing, the Vega Pro GPU in your MBP had a higher % utilisation (90%) compared to your CPU utilisation (20-50% - although clocked at 4.0GHz).
As has been previously stated in above posts - the actual testing format used by Apple has not been released, although anyone can download the Puget System's test - so it is hard to create an apples-to-apple's comparison
as 's testing may have been more GPU focused (taking advantage of the dual Vega Pros and 128GB HBM2 - 8x more HBM2 than in the iMac Pro tested [16GB HBM2]).It would be better to wait until the new Mac Pro is released and then run Puget System's benchmark first; and then maybe use a sustained heavier CPU based test to see if the differences in heatsinks between the iMac Pros and new Mac Pro (as well as a few PCs) have an influence in sustained boost CPU clocks.
Dominic
===================================================================
My background - I am a Pharmaceutical Scientist as well as being a computer hobbyiest (in my early 30s) - SoyCapitanSoyCapitan may have more IT experience than I have been alive LOL
.
Last edited:
**EDITED as I referred to the wrong poster and added more technical info**
Hi mavericks7913,
The Xeon CPU that Apple tested has yet to be released and is a generation ahead of currently available CPUs. It's boost clock is 4.4GHz, and as it is a newer generation - the Instructions Per Cycle/Clock (IPC) may be more efficient than those CPUs previously tested by Puget Systems. In other words, the New Xeon W 4.4GHz max boost could be more efficient "per clock"* than current generation 4.3GHz boost clock (iMac Pro Xeon 14 Core model has 33.25MB cache [Puget System's test], the 18 core model has 42.75MB cache [Apple's Test] - both iMac Pro models boost at 4.3GHz), and maybe even the i9 9900K (5.0GHz boost) used in Puget System's Testing quoted in Post #1.
This may be verified by Puget System's testing in the future.
*Even if the IPC were the same between the current Xeons used in the tests and new model Xeons - there is still a 0.1GHz difference favouring the upcoming generation of Xeon CPUs in the New Mac Pro, as well as higher cache (66.5MB).
The to-be-released Mac Pro also has a 300W rated heatsink - "probably" more superior
This may also be confirmed by using a sustained CPU heavy benchmark on a few Mac and PC models.
Considering that both the iMac and to-be released Mac Pro models tested has a minimum of 256GB RAM, so it could be assumed that very large files were used (a heavy work load), and that the extra RAM may have tipped things in favour of the new Mac Pro.
Apple compared the to-be-released Mac Pro with previous/current generation models on their CPU model (Post #1)
Most customers tend to identify their model by their CPU specs.
For example - People are more likely to say that they have "the 12 Core Trash Can" model rather than saying "I have the (dual) D700 trash can model^" - even if they have both the 12 CPU Cores and dual D700 GPUs in their trash can Mac Pro.
^Other trash can models with fewer CPU cores could be configured with the dual D700 graphics.
How customers identify their computer model is usually a question of context.
As a result, Apple quoted the CPU spec of each of the model tested in the graphic you provided in Post #1 instead of the GPU specs in the models used in their tests.
In theory, Apple could have stated in the graphic (attached in Post #1) that they used the following 3 models:
New Mac Pro - Dual AMD Radeon Pro Vega II graphics with Infinity Fabric Link and 32GB of HBM2 each (128GB total).
iMac Pro - Radeon Pro Vega 64X graphics with 16GB of HBM2.
Current Mac Pro - Dual AMD D700 Graphics with 6GB of vRAM/GDDR5 RAM each (12GB total).
However, as mentioned previously, most of Apple's customers would more readily identify with the CPU model used in each test than the GPU used.
It is probably correct that Photoshop would not used more than 8 cores, but the new Mac Pro had a much beefier GPU set up and more RAM (384GB compared to Puget System's 128GB RAM).
In William Payne's testing, the Vega Pro GPU in your MBP had a higher % utilisation (90%) compared to your CPU utilisation (20-50% - although clocked at 4.0GHz).
As has been previously stated in above posts - the actual testing format used by Apple has not been released, although anyone can download the Puget System's test - so it is hard to create an apples-to-apple's comparison
It would be better to wait until the new Mac Pro is released and then run Puget System's benchmark first; and then maybe use a sustained heavier CPU based test to see if the differences in heatsinks between the iMac Pros and new Mac Pro (as well as a few PCs) have an influence in sustained boost CPU clocks.
Dominic
===================================================================
My background - I am a Pharmaceutical Scientist as well as being a computer hobbyiest (in my early 30s) - SoyCapitanSoyCapitan may have more IT experience than I have been alive LOL
It is probably correct that Photoshop would not used more than 8 cores, but the new Mac Pro had a much beefier GPU set up and more RAM (384GB compared to Puget System's 128GB RAM).
https://www.google.com/url?client=i...FjAHegQIBhAC&usg=AOvVaw3mfBMy1ZlyQpN_3RuPzxvT
Photoshop does not take advantage of VRAM and they are Nvidia friendly. RTX 2060 6gb is better than Radeon VII 16gb base on the recent testing. RAM is depends on what you are doing so I would leave it as a personal preference.
The Apple bench mark clearly states " faster processing of well-threaded filters4"
and here is what the caveat supercript 4 states:
- Adobe Photoshop CC 2019 21.0.0 tested using the crystallize, pointillize, radial blur, shape blur, dust & scratches, and median filters. Performance tests are conducted using specific computer systems and reflect the approximate performance of Mac Pro and iMac Pro.
Our Photoshop benchmark (which is available for public download) tests a large number of tasks that we split into 4 categories: General, Filter, Photomerge, as well as GPU-accelerated which contains results from both the general and filter tests. With the exception of Photomerge, we also run each test in both 8 bits/channel and 16 bits/channel mode. These results all combine into a single "Overall Score" that represents the overall performance of the system in Photoshop.
And when you look at what filters Puget used compared to Apple's you will find that there are no filters in common, so you cannot use these bench marks to compare.
The only thing we know for sure is that Apple always apply a reality distortion field to all their specs in order to make them look more impressive than they actually are.
At any rate when working with Photoshop you end up mostly using brushes, creating, copying and pasting elements and layering and transforming bits, zooming, panning and rotating the canvas. You're not going to constantly run filters for e.g. producing a texture.
27 out of 28 cores will have a pretty quiet break while you're doing your thing.
Says you. I use the multithreaded filets all the time. Mostly third party ones. Better Multithreading is on Adobes roadmap throughout their apps. Taking their own sweet time about it though.
That doesnt justify using 28 cores on Photoshop. Like I said, Photoshop does not use more than 8 cores and the clock speed is the best way to increase the performance.
That statement isn't technically completely true. Some parts of Photoshop can and some larger subset of parts of Photoshop can't. The grand, sweeping assertion that "Photoshop can't" is more useful as a sales pitch for Puget than firmly grounded in truth. All of Photoshop and all of the plug-ins are not limited to 8 cores or just x86 cores.
This set of benchmarks show that GPUs do make a difference in the overall blended mix.
https://www.pugetsystems.com/labs/articles/Photoshop-CC-2019-AMD-Radeon-VII-16GB-Performance-1379/
The notion there can't be any difference past eight x86 cores isn't really true even by some of Puget's benchmarks.
I'm not sure Puget has broken down and started selling a AMD system yet but they are much solidly an Intel (and Windows ) shop. The sell Intel stuff so their benchmarks are highly likely to match what they sell. The notion that they are completely, 100% unbiased observers here is a leap. If single threaded Turbo speeds is all that tips the scales in their blended benchmarks why did they pick the iMac Pro 14 core model ( Turbo 4.3GHz) vs. the 10 Core ( Turbo 4.5GHz. ). If "beyond 8 cores doesn't matter" why did they go 4 instead of just 2 ? [ Apple has their own skew comparing just in vertical comparisons of their own product. The now ancient MP 2013 is easier to leave in the dust due to old age. ]. Puget isn't looking for benchmark elements for higher core counts because that would tip the scale toward Threadripper which they don't focus on.
On Apple's web right underneath the very large "4.2x" marker for the Photoshop speed up multiplier there is directly a note "of well-threaded filers". Apple is being very upfront here about it is a filter sped up not a "Photoshop basic core" speed up.
Puget's benchmarks are more well rounded for a broad audience with more general toolset. The Mac Pro is not being presented to the broadest audience possible. The starting price is $6K. The 28 core option uses a processor that lists for $7453 all by itself.
https://ark.intel.com/content/www/u...n-w-3275m-processor-38-5m-cache-2-50-ghz.html
Apple's mark-up is probably going to put that in the $7999 or better zone. So this system is in the $14+ K zone. That ~$7K processor is only worth more than the ~$4K processor with same cores and clocks if pushing past 1TB of RAM. [ if not going to eventually pack > 1TB of memory into this system just tossed $3K out the window right there. ] So it is only relatively high RAM ( i.e., monstrously large files and needed RAM working set ) and a narrow set of Photoshop usages where that configuration will work . Otherwise just the raw cost of the system will 'scare' mainstream photoshop users away. If knocking Apple's benchmarks it would be for picking the subset of features of Photoshop to measure that were outside what that narrow group would use. If they picked a set of benchmarks that is for that narrow set ... then they are merely talking to their target audience not the one that is running away screaming because of the basic price point.
This isn't for Photoshop work touching up High School year book photos. There are narrow set of Photoshop uses that do go off into the weeds of very custom plug. How many of those are they going to sell? Not many. Photoshop in the context of 'correcting' video end content as opposed to simple photographs, so it is really Photoshop and some other app(s) being used most of the time.
[doublepost=1560102811][/doublepost]
If folks spend 30-60% worktime using those filters then yes it does. Photoshop has a ridiculous number of features that only subsets of folks use. Mainstream Photoshop user that is only using the same features since Photoshop CS 3-4. No, it doesn't make sense. Photoshop users that are primarily using features that work around the limitation then yes it does.
Is the latter going help them sell 25K seats of 28 core systems to primarily Photoshop only users ? Highly likely not. (even over multiple years of sales. )
Last edited:
You can be sure that Apple tested it on the 28-core Mac Pro for a reason. Either those filters are already multithreaded for more than 8 cores (which would be trivial for something like the median filter, and would explain the clarification note on Apple’s website) even if most Photoshop functions aren’t, or they’ll be by the time Apple releases the Mac Pro.
That doesn’t mean that the 28-core Mac Pro is the ideal configuration for most Photoshop usage, but there’s really little point in questioning a specific benchmark of an unreleased computer on an unspecified version of the benchmarked software.
That doesn’t mean that the 28-core Mac Pro is the ideal configuration for most Photoshop usage, but there’s really little point in questioning a specific benchmark of an unreleased computer on an unspecified version of the benchmarked software.
If all you do is apply filters then that may matter more - but then wouldn't you be better off in a software that works more like Lightroom or Luminar?
As soon as you step outside filters it's going to be mostly single core speed that dictates how fast the software feels though. I wouldn't hold my breath for multithreaded interactive tools, only so much they can do to split up the workload across cores.
Similar to reasons doing it on the new iMac which has the same "well threaded" scoping for the "previous gen' comparisons there.
For the 27" model it is "70%faster processing of well-threaded filters6 " ( footnote 6 instead of 4 for this product. ) https://www.apple.com/imac/ (as of the May 2019 version of the page. ) The 2019 model's base clock drops ( 3.6 from 4.2GHz for the top BTO option) so if using core functionality that is suck at four cores it isn't going to play well.
If they are using "well threaded" on iMac on "biggest core count" it maps seemingly more consistently to do same mapping on Mac Pro. ( it is the same focused "well threaded" benchmark over at MacBook Pro also. The Fall 2018 mini ... same thing. ).
What all three have in common is that this is more cores than they every offered before so using a Photoshop subset that highlights those core count increases.
"well threaded" means that it scales with resources allocated. if look at the new iMac and Macbook Pro you'll see that going from 4 to 8 is giving in the range of ~70% improvement. So this isn't linear ( at least a slope >= 1 ). 28 cores is same: a 3.4x jump to the 18 core iMac pro and 4.2x jump to the 28 core ( ~56% jump in core and ~27% increase is 'multiple' over the MP 2013 baseline.
It is an Apple general audience sales pitch page. They are looking for a "biggest number" not some deep analysis. If look at what those additional 10 cores cost over an iMac Pro for $/performance it won't look so good. ( and they are using iMac Pro instead of iMac in the Mac Pro's chart for a couple of reasons. )
Apple is using this basic benchmark across a couple of shipping systems. It actually would be odd if there weren't using the same benchmark they have used for about the last 12 months for Photoshop on these sales pitch pages. This isn't 'new' solely for the Mac Pro. ( the new retina MacBook Air 2018 doesn't have a Photoshop benchmark but it also didn't get a core count increase either. )
Last edited:
Then you need to explain why AMD Threadripper with 32cores has a lower score than Intel i9-9900K? You better bring links or video about testing thread ripper with Photoshop. If not, I highly doubt about your statement since you are not providing any proves and information.
You said it right: in PCs. Different version of OS, different version of the software.
How exactly can you compare them?
Hell, as somebody higher up in the thread pointed out, Apple and Purget are not even testing the same thing.
I hate when people consider Purget's tests alone to evaluate the performance of a Mac.
Simply makes no sense to me.
Last edited:
If you cant believe, go and ask Adobe staffs about this topic. They will say the same thing: There isnt any difference between MacOS and Windows.
But PC will perform better than Mac because of cooling performance. Mac computers are well known for having a poor cooling performance with a small CPU and GPU fan! Do you expect to get high performance from Mac computers with a small fan? NO... This is Apple's fault who made small coolers.
These are CPU coolers for PC. Much bigger and better.
LOLOL. A small fan for both iMac and Mac mini. What are you expecting?
Last edited:
This is mostly what I was trying to say. They tested the 28-core Mac Pro for a reason: in this case, as you said, that it would show the maximum improvement. I was just trying to emphasize that the 28-core Mac Pro would show the maximum improvement because the small subset of Photoshop actions they tried are already optimized for more than 8 cores.
This is a wild statement. *Some* PCs with equivalent specs have better cooling. Some don't.
Just to put it out there Puget have tested Threadripper and do sell systems where it is a better cpu choice for the application.
https://www.pugetsystems.com/labs/articles/Photoshop-CC-2019-CPU-Roundup-Intel-vs-AMD-vs-Mac-1295/
https://www.pugetsystems.com/labs/articles/Photoshop-CC-2019-CPU-Roundup-Intel-vs-AMD-vs-Mac-1295/
At least PC have more options especially with custom PC because you can add whatever you choose while Mac is not.
If you want a PC than use a PC. At the end of the day whatever gets your work done is what’s best for you.
The tool doesn’t matter as long as it does the work.
lol stop trolling. Do you even aware of how many Adobe users are using Mac? Mac and Windows are totally different in terms of OS and yet you are asking me to get PC. How ignorant you are. We have freedom of speech to criticize Mac computer's poor cooling system. It is a fact that Apple ruined their products just because of the cooling system such as Apple III, PowerMac G4 Cube, and Mac Pro 2013. You didnt learn anything from their past.
If you dont wanna hear the critism toward Mac's cooling system, then leave this thread.
[doublepost=1560116806][/doublepost]
And yet the performance is way lower than Intel i9-9900K
Yes I know many. Myself included. We have no control over cooling. Cooking can always be better but that is something that Apple have to fix. We as users have no control over that.