RAM Usage in Vectorworks 2022 SP4.1

[erminio]

Hi all guys
In the last Vectorworks release note known problems about RAM usage in VP sections are declared to be solved.
As I can see things are maybe a little bit better now, but not solved at all
I usually work with files of even 500MB, containing complex DWGs and DTMs
I'm still looking at Activity monitor, trying to figure out the best moment to quit Vectorworks before it stops working by itself
Do someone else still experience this problem?
Suggestions about that?

 

My computer is a iMac 27 2017 i7 4,2GHz of RAM and 1TB NvMe with Monterey 12.5

 

I've seen in Vectorworks 2023 NNA promises to reduce RAM usage by 80% and speed up VP Sections to 6 times
At the same time I've also seen that specs requests for the real world High End profile are quite important (Mac Studio Max with with 64GB or better), so this confuses me quite a bit, since I expected RAM needs for silicons to be lower than Intel PCs

 

So can someone confirm that in next (if any) 2022 SP or at least in VW 2023 things will improve with RAM?

 

PS: I'm planning to buy an M1 Max 32 core GPU with 64GB of RAM and 1TB of disk, but since this configuration is practically just the high end profile base model 😥, I'm afraid it could become old in a couple of years (M1 ultra is quite expensive for me)

Suggestions?

Thank you for your attention

[Jeff Prince]

6 hours ago, erminio said:

 

PS: I'm planning to buy an M1 Max 32 core GPU with 64GB of RAM and 1TB of disk, but since this configuration is practically just the high end profile base model 😥, I'm afraid it could become old in a couple of years (M1 ultra is quite expensive for me)

Suggestions?

I don’t know the answer to your question, but I can tell you the last Intel IMac w/8GB vram & 72GB RAM meets their high profile spec and I wouldn’t by an M1 yet.

 

When comparing the Studio M1Max vs the IMac with VWX 2022, site model updates, on a complex example, were barely faster on the M1 based machine, which was a major disappointment.  Even comparing the same operations to the 2015 MBP failed to deliver impressive results.  In all three machines, complex site models will pin a core of the processor and indicate that the program is not responding, you just wait for it to finish and hope you have enough RAM.

 

The benchmarks all say the newer computers should be far faster, but if the VWX is only running on one core there seems to be a major issue.  I have been told many reasons why this is the case, but I suspect Apple and vectorworks will deliver better performance by early next year.

 

People rendering and just wondering around the OS on the new macs have wonderful things to say, I haven’t heard anyone bragging about site model edits improving.  Perhaps this is the best multi core vs single core real world example to consider.

 

 

[zoomer]

"RAM Usage in Vectorworks 2022 SP4.1"

 

Do we have a new (sub) Service Pack for VW 2022 US ?

Or are you using the German or other localized Version ?

 

I also had larger problems and crashes with SLVP updating

or even any DLVP editing - pre SP4.

At least the memory leeks seem to be solved with SP4.

 

But I am not sure on the overall speed of updating SLVPs

as I have no more comparison.

(But it feels a bit unrealistically slow though)

Edited August 18, 2022 by zoomer

[erminio]

On 8/18/2022 at 4:20 PM, jeff prince said:

I don’t know the answer to your question, but I can tell you the last Intel IMac w/8GB vram & 72GB RAM meets their high profile spec and I wouldn’t by an M1 yet.

 

When comparing the Studio M1Max vs the IMac with VWX 2022, site model updates, on a complex example, were barely faster on the M1 based machine, which was a major disappointment.  Even comparing the same operations to the 2015 MBP failed to deliver impressive results.  In all three machines, complex site models will pin a core of the processor and indicate that the program is not responding, you just wait for it to finish and hope you have enough RAM.

 

The benchmarks all say the newer computers should be far faster, but if the VWX is only running on one core there seems to be a major issue.  I have been told many reasons why this is the case, but I suspect Apple and vectorworks will deliver better performance by early next year.

 

People rendering and just wondering around the OS on the new macs have wonderful things to say, I haven’t heard anyone bragging about site model edits improving.  Perhaps this is the best multi core vs single core real world example to consider.

 

 

Thank you Jeff
Very useful

[erminio]

On 8/18/2022 at 9:55 PM, zoomer said:

"RAM Usage in Vectorworks 2022 SP4.1"

 

Do we have a new (sub) Service Pack for VW 2022 US ?

Or are you using the German or other localized Version ?

 

I also had larger problems and crashes with SLVP updating

or even any DLVP editing - pre SP4.

At least the memory leeks seem to be solved with SP4.

 

But I am not sure on the overall speed of updating SLVPs

as I have no more comparison.

(But it feels a bit unrealistically slow though)

Sorry

I didn't specify I'm using Italian Version

Didn't think it could be important

In Ita version memory leeks don't seem to have been resolved yet

Anyway it seems a SP5 is arriving 

 

Thank you Zoomer for your interest

[zoomer]

6 hours ago, erminio said:

I didn't specify I'm using Italian Version

 

OK, localized versions need some more 2 weeks or so to offer the SP

and usually integrate some fixes for their own Tools.

So they call it e.g SP4.1.

But I think the core changes, like fixing some memory leeks are in caharge

of NNA and so your SP4.1 has just SP4 fixes.

If VW NNA would have released a SP4.1 already one could hope that there

were some further core fixes included.

 

As I said, SP2 (?) to SP3 looked totally buggy for VP editing and it got much

better with SP4 for me personally on my small M1 Mac Mini.

 

But I also read that the complaints on the forum did not totally stop with SP4.

So there may still be some issues left.

I just have not intensively tested against VW 2021 or earlier versionsto be sure

they have fixed everything. E.g. RW VP renderings for me still "feels" unrealistically

time consuming but therefore, that is just a feeling, no real evidence.

 

[zoomer]

On 8/18/2022 at 4:20 PM, jeff prince said:

When comparing the Studio M1Max vs the IMac with VWX 2022, site model updates, on a complex example, were barely faster on the M1 based machine, which was a major disappointment.

 

I am not sure.

Yes, some of VW CAD tasks are single tasked and may need to stay ST forever.

But as far as I get it, especially even the lowest M1 Models single core

performance was pretty impressing.

Of course much faster than my previous Trash Can Xeon and still faster

than my not-shy of single core Ryzen 3950X. Of course latest Intels or

Ryzens got a bit faster in SC performance than Apple ARM M1 family,

(basically all at the same single core speed)

M1s are still in the top ten of the CPU single core benchmarks list.

 

I do not notice much difference between 3950X on PC vs my M1 Mini,

but 2013 Trash Can Xeon vs M1 Mini is more than noticeable and more

fluid anywhere.

16 GB on Apple ARM is of course far to less for larger projects, but

they do not run fine on unlimited PC although anyway and most

mid size projects run at least as fast on my M1.

 

[erminio]

On 8/18/2022 at 4:20 PM, jeff prince said:

 

The benchmarks all say the newer computers should be far faster, but if the VWX is only running on one core there seems to be a major issue.  I have been told many reasons why this is the case, but I suspect Apple and vectorworks will deliver better performance by early next year.

 

 

After I've read your post I've given a look to Activity Monitor while a SLVP calculation of a complex model with DTM and I've seen Vectorworks uses half cores and never at max speed, even if the calculation takes a while

So it seems to be multicore

Maybe it has to be optimized, don't know

 

[erminio]

On 8/22/2022 at 9:47 PM, zoomer said:

 

OK, localized versions need some more 2 weeks or so to offer the SP

and usually integrate some fixes for their own Tools.

So they call it e.g SP4.1.

But I think the core changes, like fixing some memory leeks are in caharge

of NNA and so your SP4.1 has just SP4 fixes.

If VW NNA would have released a SP4.1 already one could hope that there

were some further core fixes included.

 

As I said, SP2 (?) to SP3 looked totally buggy for VP editing and it got much

better with SP4 for me personally on my small M1 Mac Mini.

 

But I also read that the complaints on the forum did not totally stop with SP4.

So there may still be some issues left.

I just have not intensively tested against VW 2021 or earlier versionsto be sure

they have fixed everything. E.g. RW VP renderings for me still "feels" unrealistically

time consuming but therefore, that is just a feeling, no real evidence.

 

So Italian SP5 has finally arrived, but in my case didn't resolve memory leeks at all

When launching complex SLVPs with DTMs Vectorworks takes all memory and never gives it back till I quit 

Really don't know what to do...  

 

Should be nice a sort of "Empty cache" command
Don't know if it is possible...

 

[ashot]

On 9/6/2022 at 1:19 AM, erminio said:

So Italian SP5 has finally arrived, but in my case didn't resolve memory leeks at all

When launching complex SLVPs with DTMs Vectorworks takes all memory and never gives it back till I quit 

Really don't know what to do...  

 

Should be nice a sort of "Empty cache" command
Don't know if it is possible...

 

Hi Erminio, can you please tell me what utility you used fo rmemory usage shown on screenshot?

 

[erminio]

12 hours ago, ashot said:

Hi Erminio, can you please tell me what utility you used fo rmemory usage shown on screenshot?

 

Hi
Sure

It's iStatistica

You can find it on the web or in the Apple Store

1st one is more complete but more expensive