Is it any good for High end Photoshop use (commercial retouching)

[pierre1610]

I'm seeing lots of benchmarks for video work, rendering work and some audio work but not much in relation to Photoshop. As a commercial retoucher i rarely use any of the gpu related features of photoshop i'd like to see test on:
large file handling 4-10GB (opening, saving of 8bit, 16bit and 32bit images)
converting large files from 8bit-16bit

I'd also like to know how it deals with rosetta Photoshop since UXP doesn't have all the functions required to re-write a lot of scripts used in our workflow

Has anyone seen anything or can anyone run some tests from PugetBench and macperformanceguide diglloyd tools.
My current systems have 64GB and 128GB ram so i'd be interested to see how unified memory compares. The Ultra seems overkill for photoshop however its the only way to get a lot of ram.

 

[davys]

I'm seeing lots of benchmarks for video work, rendering work and some audio work but not much in relation to Photoshop. As a commercial retoucher i rarely use any of the gpu related features of photoshop i'd like to see test on:
large file handling 4-10GB (opening, saving of 8bit, 16bit and 32bit images)
converting large files from 8bit-16bit

I'd also like to know how it deals with rosetta Photoshop since UXP doesn't have all the functions required to re-write a lot of scripts used in our workflow

Has anyone seen anything or can anyone run some tests from PugetBench and macperformanceguide diglloyd tools.
My current systems have 64GB and 128GB ram so i'd be interested to see how unified memory compares. The Ultra seems overkill for photoshop however its the only way to get a lot of ram.

Pierre, have a look at ArtisRight on YouTube - he’s done a wonderful series of benchmark tests on the new Studio Max base specifically for pro photographers, including diglloyd.

 

[pierre1610]

thanks, some strange results there, it seems the laptops run better.

 

[davys]

thanks, some strange results there, it seems the laptops run better.

Yes, very surprising results in that the 14" and 16" Max and Pros are often that little bit faster than the new base Studio Max in most of the tests, albeit we're often just talking about 20 seconds or so.

Given that I've really got to move on from my 2014 27" iMac at some point over the next few months, I'll probably end up going for the Studio Max with 2TB storage, as I'll still have to pay for a new monitor as well. My mobile/travel laptop needs are still being met by my 2019 (intel) 16" MacBook pro.

 

[pierre1610]

For 3.5-5.7k you'd expect a flyer!

i'm still running Mac Pro's 2010 (maxed out with pcie SSD's) and Mac Pro 2013 so i'm hoping for a big jump.

My top spec 2018 Mac Mini is faster than both but not by much.

 

[AlteMac]

For 3.5-5.7k you'd expect a flyer!

i'm still running Mac Pro's 2010 (maxed out with pcie SSD's) and Mac Pro 2013 so i'm hoping for a big jump.

My top spec 2018 Mac Mini is faster than both but not by much.

FWIW, I think you'll see a huge jump. I saw a quite noticeable jump in PSCC and Nik and Topaz plugins going from a 2013 MP with 64G to a Mac Mini M1 with 16G. For my other 2013 MP I am switching to Max Studio/64G.

 

[Chancha]

With your PC, are you seeing significant improvement when having 128GB vs 64GB? On the Apple Silicon Macs since the same memory pool is shared for VRAM as well therefore some GBs have to be discounted from there for calculation.

So far only very few M1 Ultra tests have been available concerning photographic tasks, and almost all are GPU 48 cores stock model not the topped out 64 cores.

 

[fisherman188]

Happy to run any tests you want on a base model Max Studio.

I run lightroom and Photoshop pretty extensively but my files aren't quite as large as some of the ones you're working with.

Lightroom moves through the catalog and applies settings better than I've ever seen.

In Photoshop, When zoomed in on a 100MP panorama looking for dust spots or hot pixels, it's buttery smooth as I drag across (never had it them smooth before).

Topaz Denoise and Sharpen are crazy! On my 2018 i7 2.7Ghz MBP, sometimes the tools would take 1-2 minutes to update a view on a image with lots of edges. On the Studio updates take under 2 seconds for 25MP-ish images and most are under 1 second. Editing is finally enjoyable. I haven't been able to use more than 26GB RAM with LR, PS, and a Topaz plugin open so I'm not sure the 64GB RAM would be of much help to me.

I'm incredibly happy with the performance but probably returning this model for a 1-2TB SSD when they're available.

 

[pierre1610]

Happy to run any tests you want on a base model Max Studio.

I run lightroom and Photoshop pretty extensively but my files aren't quite as large as some of the ones you're working with.

Lightroom moves through the catalog and applies settings better than I've ever seen.

In Photoshop, When zoomed in on a 100MP panorama looking for dust spots or hot pixels, it's buttery smooth as I drag across (never had it them smooth before).

Topaz Denoise and Sharpen are crazy! On my 2018 i7 2.7Ghz MBP, sometimes the tools would take 1-2 minutes to update a view on a image with lots of edges. On the Studio updates take under 2 seconds for 25MP-ish images and most are under 1 second. Editing is finally enjoyable. I haven't been able to use more than 26GB RAM with LR, PS, and a Topaz plugin open so I'm not sure the 64GB RAM would be of much help to me.

I'm incredibly happy with the performance but probably returning this model for a 1-2TB SSD when they're available.

I'll see if i can find something to test, for now could you run the diglloyd benchmarks pls

diglloyd Benchmarks for Photoshop CC

macperformanceguide.com

Specifically the diglloydSpeed1 and diglloydHuge tests. I've seen results on youtube but they seem way off especially the diglloydHuge result, its slower than my 2010 Mac Pro result

I know the apple silicon version of Photoshop killed off CEP panels. Do you know if you can run .jsx scripts though?

 

[fisherman188]

I'll see if i can find something to test, for now could you run the diglloyd benchmarks pls

diglloyd Benchmarks for Photoshop CC

macperformanceguide.com

Specifically the diglloydSpeed1 and diglloydHuge tests. I've seen results on youtube but they seem way off especially the diglloydHuge result, its slower than my 2010 Mac Pro result

I know the apple silicon version of Photoshop killed off CEP panels. Do you know if you can run .jsx scripts though?

I'm getting this error when I try to run either test, is there something I need to set first?

 

[pierre1610]

View attachment 1980066

I'm getting this error when I try to run either test, is there something I need to set first?

you have to run the tests from the "runDiglloydBenchmark-v4_2_3" app within the same zip file, not the actions individually in photoshop

 

[fisherman188]

I'm running the App straight from the Zip

 

[OSB]

With your PC, are you seeing significant improvement when having 128GB vs 64GB? On the Apple Silicon Macs since the same memory pool is shared for VRAM as well therefore some GBs have to be discounted from there for calculation.

That's not really the way that UMA works on the M1. The GPU doesn't carve out a separate pool of memory and hold onto it as dedicated VRAM. It's all the same memory, and if data is in-memory, both the CPU cores and the GPU cores can act upon it without moving it about.

 

[now i see it]

Is it any good for High end Photoshop use ?(commercial retouching)​

The Mac Studio (any version) is one of the fastest personal computers on earth.
To put things in perspective, all the best photoshop jockeys in the world used computers leagues slower than this years ago.

 

[pierre1610]

I'm running the App straight from the Zip
View attachment 1980165

Have to installed the atn file into photoshop too? Drag the .atn file onto the photoshop icon in the dock to do this. Then run the bechmark as you have been.

 

[pierre1610]

Is it any good for High end Photoshop use ?(commercial retouching)​

The Mac Studio (any version) is one of the fastest personal computers on earth.
To put things in perspective, all the best photoshop jockeys in the world used computers leagues slower than this years ago.

Yes, i've been using Photoshop since the days of the G5. I have a PC built in 2020 which is no faster than a Mac built in 2010. This is why i want to know if the Studio is worth the investment.

 

[fisherman188]

Alright, diglloyd Photoshop tests on base model M1 Max Studio:

Speed1
View attachment 1980453

Huge
View attachment 1980517

In the middle of "huge" test
View attachment 1980460

 

[pierre1610]

Thanks for doing that. Whats the spec? 10CPU - 24GPU - 32GB Ram?
It seems your huge test was a little faster than the youtube vid by 7 secs. Pretty similar really though.

Anyone got an Ultra? Be interested to see how the 64GB and 128GB versions compare.

Thanks

 

[fisherman188]

Thanks for doing that. Whats the spec? 10CPU - 24GPU - 32GB Ram?
It seems your huge test was a little faster than the youtube vid by 7 secs. Pretty similar really though.

Anyone got an Ultra? Be interested to see how the 64GB and 128GB versions compare.

Thanks

Yup.

M1 Max 10 CPU, 24, GPU, 16 NC
32GB RAM
512GB SSD

Planning to return it next week and get a 1TB or 2TB model when the stock comes back. I'm happy with its performance which is a huge jump from my 2018 MBP. Probably going to move to a move to a Studio at home then pick up an M2 Air for when I'm traveling. There's just something nicer about sitting down to a permanent desktop instead of always having to plug in the laptop. But there is still the option of an M1 Max 16 MBP that I'm also considering since the Max seems to be plenty power in LR and PS. Was really hoping that huge CPU heatsink and fan were going to have the Max running faster in the Studio than the MBP but looks like it's just over engineered for now.

 

[pierre1610]

For anyone interested this is what we're seeing.
Diglloyd Huge results:
2022 MacStudio Max 10CPU 32GB RAM = 73.7
2021 MBP M1 MAX 64GB RAM = 19
2019 MacPro 28 Core 384GB RAM = 37.1
2019 iMac 5k 8-core 128GB RAM = 23.8
2018 MacMini i7 64GB RAM = 40.89
2013 MacPro 6-core 128GB RAM = 32.65
2010 Mac Pro 3.33GHz 32GB = 172 (this is probably much higher with modern PCIe SSDs)

Photoshop really is a Ram whore,
2013 MacPro 6-core 64GB RAM = 46.37

The M1 MBP scores the best with 64GB RAM so i'm hoping the MS 64GB will score the same, im surprised the 32GB MS is so low, all i can thing is it needs more ram.

Here's hoping the 128GB MacStudio can score around 15.

 

[lcubed]

for context, are larger scores or lower scores better for digilloyd huge?

 

[pierre1610]

Lower scores are better

for context, are larger scores or lower scores better for digilloyd huge?

 

[whwang]

Looks like this is a test of how much RAM you have, rather than how fast your system is.

 

[fisherman188]

Looks like this is a test of how much RAM you have, rather than how fast your system is.

I wouldn't say so. It only hit a max of 27GB or RAM on my 32GB machine.

 

[whwang]

I wouldn't say so. It only hit a max of 27GB or RAM on my 32GB machine.

I would say this hits the limit. The 32GB of RAM is shared with system, Photoshop, and perhaps some other apps running in the background. It's possible that the test requires way more than 27GB, so it's swapping your SSD when the system cannot allocate more RAM to it.

If the max is 27GB in a 64GB machine, then I can agree that this test really only needs 27GB.