Jump to content

Mac Studio & M1 Ultra


Recommended Posts

2 hours ago, Sky said:

The Apple Studio Display is 5K in a 27" package.

Does that mean the VW UI elements would get smaller?

Or are those dynamically resized depending on the monitor?

 

Yes but you have options (literally; press the Option key when you hit the Scaled button) with all monitors:

 

676338463_ScreenShot2022-03-09at1_55_32PM.thumb.png.99d0da0f6445ffa9033b92b9e207c914.png

  • Like 1
Link to comment
3 hours ago, J. Miller said:

Mark,

Can you tell me what monitors you are using. I also sit at the desk full time now-a-days.

speakers - would be nice , video cam - yes, 4k - yes, 27" - min width.

I am planning on getting (2)

Jeff

 

From experience:

  • Dell U4021QW 40" Curved
  • BenQ PD3200U 32"

Also look good:

  • Dell U4320Q 42.5"
  • BenQ PD3220U DesignVue 31.5"

I love having 1 Thunderbolt cable connected to my MBP from the monitor (using the monitor as a dock / hub) although that's not really as beneficial with the Studio.

 

Curved is a nice-to-have (when you hit 40" and above) but not a must-have.

 

The must-have is vertical pixels. You're going to use the 2nd highest resolution (not the max res advertised on the box). As a rule of thumb:

  • 27" = 1440
  • 32" = 1620
  • 43" = 2160 (full 4K UHD)

Most widescreen monitors top out at 1440 (capable of 1600 but microscopic in real life), so they're useless to me.

 

I have a Logitech webcam.

 

Loads of great speaker options if you wanna go down that rabbit hole... 

 

Parting thought: 32 is the new 27 (shame on you, Apple).

 

 

Edited by Mark Aceto
Link to comment
37 minutes ago, Mark Aceto said:

Yes but you have options (literally; press the Option key when you hit the Scaled button) with all monitors:

 

 

Usually in automatic mode, Apple will scale the Monitor to 200%.

So you will get about the same space for Palettes like a standard 27"

Monitor with 2560x1440 resolution.

But it will be double as sharp.

 

Like for my 24" 4K Monitor on Windows it will recommend a 200% scaling.

As it is so sharp, I allow myself to scale at only 175%, to get some more

screen space for Palettes.

 

You can do the same on Mac with manual scaling.

Just that uglier numbers than full 100%, 200%, ... scaling will cost a bit of

crispness.

  • Like 1
Link to comment

It's worth putting some thought into resolution with external displays - there are a few things to consider, especially the relationship between monitor size and pixel count because it some implications for how things get scaled.

 

I found this article very useful, especially the table which gives you a way of seeing which combinations fall into what they call the "bad zone".

 

https://bjango.com/articles/macexternaldisplays/

Link to comment
3 hours ago, Mark Aceto said:

Parting thought: 32 is the new 27 (shame on you, Apple).


Is it really though?  There comes a point where big, is in fact, too big for monitors IMHO.  I’m productive with a 16” mbp, super happy on a 27” IMac, feel like I’m sitting too close to the movie screen on anything larger 🙂

  • Like 1
Link to comment
12 hours ago, Sky said:

The Apple Studio Display is 5K in a 27" package.

Does that mean the VW UI elements would get smaller?

Or are those dynamically resized depending on the monitor?

 

I used to use a 27" 5K iMac and have since moved on to a 32" 6K Pro Display XDR and to my eyes the scaling works perfectly. Icons and GUI are all a very nice size. One thing to keep in mind though is that more pixels probably use more computer power and my MacBook Air is probably on the short side for the 6K.

Edited by Anders Blomberg
  • Like 1
Link to comment
1 minute ago, Anders Blomberg said:

 

Thanks for engaging in this @SteveJ! It's nice that Vectorworks is taking advantage of multiple CPUs in some tasks but from my experience the typical CPU load looks as below. The task that triggered this single core spike was changing the name of a landscape ares style, reflected in roughly 120 landscape area objects. While VW might be able to utilise multiple cores for many tasks I see these single core spikes for so many tasks in my day to day use. It is a little frustrating having to sit and wait for tasks to compute and at the same time seeing all those other cores not being utilised. I'm glad for the improvements VW have done in this area and look forward to further improvements!

 

image.png.6921a5dbd9bb812be7d80b47bb037383.png

 

That's why M1 ultra show 1,6x and M1 max 1,3x?

 

731224309_Schermata2022-03-10alle08_42_08.thumb.jpg.f9197ae36ea69abceb380aaa4a0084b2.jpg

Link to comment
22 hours ago, Anders Blomberg said:

The performance increases for other software are reported to be 3x or 4x but for VW it peaks at 1,6. I'm guessing this is because performance increases for processors these days seem to come mainly from more cores, and Vectorworks only use a single core in many tasks, as I understand it.

 

Absolutely agree.

  • Like 1
Link to comment
11 hours ago, zoomer said:

Just that uglier numbers than full 100%, 200%, ... scaling will cost a bit of

crispness.

 

 

I forgot something,

today there are still lots of Apps that are not HiDPI ready.

That means, as soon as you need Display Scaling you can get problems with GUI.

Like some hard coded Panel sizes and Fonts that will not match or react to scaling.

 

 

In Bricscad on Linux, when using e.g. my default scaling of 175%, it does not scale

the fixed Start Panel, but its content, so you can't reach the hidden control elements

to even start the App. Bricscad can only use a 200% Scale because of the Icons.

 

  • Like 1
Link to comment

I'd like to give a couple of examples of things where I have found the performance of Vectorworks 2022 on M1 disappointing.

 

These are both "everyday" tasks which, having been told that VW2022 has native support for Apple Silicon, I might have hoped would see improved performance compared to VW2021 on the same computer. But in fact what I see is either no improvement or actually *worse* performance.

 

They are both tasks that I might be commonly carrying out in a repetitive manner and therefore have a significant impact on speed of working. And although this is a large and complex file, I don't see that the tasks themselves are ones that ought to be computationally difficult.

 

This is VW2022 SP2.1 on an M1 mac mini with 16GB RAM. The VWX file size is around 2GB.

 

First, moving a single viewport on a sheet layer, then undoing the action:

 

 

 

Second, double clicking on a group to edit within that group, then double clicking a symbol to edit the symbol, then exiting the symbol, and exiting the group. This doesn't happen as painfully slowly as the previous example, but in real world use it's something you might do repeatedly and it all adds up. To me it doesn't feel unreasonable to expect that something like this (entering and existing a container object) should happen instantaneously.

 

 

Someone may tell me that this is all down to lack of memory.

 

You can see that VW is occupying a lot of memory - about 30GB, which in my limited understanding means it's swapping a lot of stuff in and out of some kind of cache somewhere. But why is it using so much memory when the exact same file in VW2021 seems to use up about half that amount?

 

I meet the "mid level profile" hardware requirements listed here

https://www.vectorworks.net/sysreq

although I don't meet the "high level profile".

 

Therefore I might expect that I am pushing things a bit with this file which is quite complex, so I would forgive things being a bit sluggish if I want to navigate the whole model in 3d with shadows on, or something like that. But these seem quite simple tasks to me. Or is it that the file size is simply one that an M1 machine with 16GB RAM can't be expected to handle efficiently?

  • Like 2
Link to comment

I experience the lags with DLVP actions too.

(I thought that would be a general cross platform VW 2022 feature !?)

 

And yes, as soon as memory pressure gets browny, it is no more fun to work on M1.

I am pretty sure this will be better with 32 or 64 GB memory.

Link to comment
On 3/9/2022 at 2:19 AM, Anders Blomberg said:

The performance increases for other software are reported to be 3x or 4x but for VW it peaks at 1,6.

 

I would not take Apple's marketing literally at all (when comparing VW improvements to other other software improvements). Their graphs and charts are notoriously misleading.

  • Like 1
Link to comment
8 hours ago, line-weight said:

M1 mac mini with 16GB RAM. The VWX file size is around 2GB.

 

That's the issue. And you're not alone. Everyone I work with that uses those first gen M1 Airs and Mini's is in the same boat. The issue isn't the CPU. It's the GPU and RAM.

 

VW22 is compatible with ARM but it needs more RAM and GPU than those systems provide. Especially with a 2GB file.

Link to comment

For the 16gb RAM constraint, think of it like packing a truck. If you have 70' of gear, you'll need 2x 53' trucks. You may only need 20gb of RAM for a task but the containers are 16 and 32, so it's not necessarily that you need all 32.

 

Also, if Redshift or Twinmotion or other apps or processes are gobbling up 8gb of VRAM--or more--of shared memory, subtract that from your 16/32/48/64/128.

  • Like 1
Link to comment
8 hours ago, line-weight said:

I'd like to give a couple of examples of things where I have found the performance of Vectorworks 2022 on M1 disappointing.

 

These are both "everyday" tasks which, having been told that VW2022 has native support for Apple Silicon, I might have hoped would see improved performance compared to VW2021 on the same computer. But in fact what I see is either no improvement or actually *worse* performance.

 

They are both tasks that I might be commonly carrying out in a repetitive manner and therefore have a significant impact on speed of working. And although this is a large and complex file, I don't see that the tasks themselves are ones that ought to be computationally difficult.

 

This is VW2022 SP2.1 on an M1 mac mini with 16GB RAM. The VWX file size is around 2GB.

 

First, moving a single viewport on a sheet layer, then undoing the action:

 

 

 

Second, double clicking on a group to edit within that group, then double clicking a symbol to edit the symbol, then exiting the symbol, and exiting the group. This doesn't happen as painfully slowly as the previous example, but in real world use it's something you might do repeatedly and it all adds up. To me it doesn't feel unreasonable to expect that something like this (entering and existing a container object) should happen instantaneously.

 

 

Someone may tell me that this is all down to lack of memory.

 

You can see that VW is occupying a lot of memory - about 30GB, which in my limited understanding means it's swapping a lot of stuff in and out of some kind of cache somewhere. But why is it using so much memory when the exact same file in VW2021 seems to use up about half that amount?

 

I meet the "mid level profile" hardware requirements listed here

https://www.vectorworks.net/sysreq

although I don't meet the "high level profile".

 

Therefore I might expect that I am pushing things a bit with this file which is quite complex, so I would forgive things being a bit sluggish if I want to navigate the whole model in 3d with shadows on, or something like that. But these seem quite simple tasks to me. Or is it that the file size is simply one that an M1 machine with 16GB RAM can't be expected to handle efficiently?

 

Hi

 

what happened on CPU tab on same process?

Link to comment

So Mark:

 

Interesting conversation. As noted by several (and experience over the years) it seems the central issue is the inability of the software to keep up with advancements in hardware (remember the days when it was the other way around?). NNA was slow to exploiting advancements in GPU processing. I agree that it appears to users that the Semens kernel is a leaking CPU hog. Reading a lot of threads and based on my daily Vectorworks crashing the entire OS, 2022 is an unmitigated graphics disaster. Apparently 32gb of ram, 10 cores, and 16GB of gpu ain't enough.  I've stopped rendering in Vectorworks now that the hare that is twinmotion beats the pants off the Vectorworks turtle.

 

But I digress. Not sure 2022 would take full advantage of a $ 6,000 AMD processor. But I drink the Apple kool-aid. So to your point, to Ultra or not. Clearly 64gb ram or the highway, but is the extra $1K for 64 core gpu worth it when renderworks uses what, 8 cores? Can't remember, do I have that right? How many banks am I robbing? My vintage 2017 iMac Pro xeon 3 ghz, 10 core benches out at 11,463 while the M1 max 10 core 3.2 rolls in at 12,402. Pretty much no difference. Ultra on the other hand, doubles the score along with the price. The trade-in however helps soften the blow or at least covers the cost of the fancy dancy $1500 monitor.

 

Thoughts?

-J

Link to comment
22 hours ago, Mark Aceto said:

Ugh, early Cinebench scores not looking good... 

 

https://www.cpu-monkey.com/en/cpu_benchmark-cinebench_r23_multi_core-16

 

1441725054_ScreenShot2022-03-09at1_43_57PM.thumb.png.97e729d9fa1bfe01e7247878c7ad7487.png

 

Much further down the list, are the M1 Max chipsets, scaling predictably at 1/2 the Ultra scores:

 

 

875393728_ScreenShot2022-03-09at1_47_44PM.thumb.png.630d80ec34ac931171d5715b5f2631a8.png

 

So, it stands to reason, that the imminent Mac Pro with 4x Max (double Ultra) would theoretically scored 48,000 (not even close to Threadripper).

 

So what's this about?

https://www.tomshardware.com/news/apple-m1-ultra-nearly-matches-amd-threadripper-3990x

 

Edit: oh it's Geekbench

Edited by Christiaan
  • Like 2
Link to comment
1 hour ago, jnr said:

nteresting conversation. As noted by several (and experience over the years) it seems the central issue is the inability of the software to keep up with advancements in hardware (remember the days when it was the other way around?).

 

VW and Maxon were one of the first to adopt native M1. I could list a litany of 3D design apps that don't even support macOS.

 

1 hour ago, jnr said:

Reading a lot of threads and based on my daily Vectorworks crashing the entire OS, 2022 is an unmitigated graphics disaster.

 

I almost never have crashes. However, I have multiple Force Quits daily. I had the same iMac Pro as you in 2018 but sold it in 2019. It's a great machine. More than capable for VW. I never once pegged the GPU on it. Then again, I've never pegged the GPU on a 2018 MBP or this 2019 MBP.

 

1 hour ago, jnr said:

 I've stopped rendering in Vectorworks now that the hare that is twinmotion beats the pants off the Vectorworks turtle.

 

TM/UE and any other realtime GPU-biased renderers will peg any GPU you throw at it, so that should answer your next question about spec's. Then again, hopefully Apple will release hardware accelerators so Mac users can use Path Tracer. And hopefully Carbon for Unreal will release a Mac plugin.

 

Twinmotion system requirements for Mac:

 

Minimum Recommended requirements 

Use case: Small to average project, ( <1GB geometry data ), 
Suitable for real-time 3D presentation, video generation in HD mode, all images formats  up to 4K
Not suitable for VR / 4K video rendering /  360 panoramic video

Operating System: Mac OS X 10.14.6 or later
Graphic Card: 6 GB dedicated memory or card with benchmark score over 10 000 here
CPU (processor): CPU with a benchmark score of 2000 higher here
System Memory (RAM): 32 GB or more
Hard Drive Space: 30 GB of free disk space

 

High-end Requirements

Use case: Large project, Airport, Large building, City, large landscape (>1 GB geometry data) 
Suitable for real-time 3D presentation, video generation in any resolution up to 8K,  360 panoramic videos, all images formats.

Operating System: Mac OS X 10.14.6 or later
Graphic Card: 12 GB (or more)  dedicated memory or card with benchmark score over 16 000here 
CPU (processor): CPU with benchmark score of 2500 higher here
System Memory (RAM): 64 GB or more
Hard Drive space: 30 GB of free disk space
Hard drives like SSD (Solid State Drive) or HDD (Hard Disc Drive) should also be taken into consideration when working a very large project. 
Storing your project on a SSD will allow your project to load and save faster.

Note: Path Tracer is not supported.

 

————————————————————

 

The interesting thing to me is that we have hindsight now compared to the announcement 2 years ago. We know that the first M1 has roughly the same single core clock speed as the M1 Pro / Max / Ultra:

 

974222609_ScreenShot2022-03-10at12_29_03PM.png.91fe11441bbf1fac7304ca177e6497f6.png

 

Looking at the Cinebench scores (useful for Renderworks) :

  • M1 8-core = 9,569
  • M1 Pro 10-core = 12,390
  • M1 Max 10-core = 12,402
  • M1 Ultra 20 core = 23,566

For comparison:

  • Threadripper 3960x 24-core = 34,255
  • Threadripper 3970x 32-core = 47,102
  • Threadripper 3990x 64-core = 74,422

So, the past 2 years seem to indicate that the last M1 to be announced, the Mac Pro, will score roughly 48,000 range (4x M1 Max). In other words, comparable with the 32-core Threadripper 3970x which users can also run all manner of 3D apps, and Path Tracer, that aren't Mac-compatible for roughly $7,500. I'm assuming the Mac Pro will cost double that for a similarly spec'd system. We'll find out in a few months...

 

image.thumb.png.4a7f5bc2d778053deebbef6998cdcad8.png

 

But if you don't care about Renderworks or Cinema, those Cinebench scores are irrelevant. Realtime render machines are all about GPU, VRAM, RAM.

 

Then there's the question of the roadmap: M2, M3... what to expect in 2023, 2024... 

 

One possible timeline is that Apple finally catches up to AMD, and Windows-only developers finally decide to bring native Mac apps. May as well lump AAA games and flying cars into this fantasy. But the reality is AMD and Intel are motivated now... 

 

Personally, I'm on the fence. I ordered a Studio yesterday but the ship day is in May. I may cancel it, and just get a Threadripper workstation. MacBook Pro as a daily driver, and the Windows workstation for heavy lifting. I'm coming down from the hype of yesterday's reality distortion field (the Kool-Aid gets me every time).

 

One thing's for sure though: that overpriced monitor is a joke.

 

 

Edited by Mark Aceto
Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...