Teaser Tuesday - Vectorworks Graphic Module for Sheet Layers - Vectorworks 2019

[PVA - Admin]

1 hour ago, line-weight said:

Is this being dealt with, with a sense of urgency? It needs to be. (Please)

Yes, we are tracking it as VB-145786. Originally it went in as a bug against Push/Pull and BIMObjects specifically, but on further investigation it seems more likely that it's not BIMObjects specifically, but the fact that they're normally made up of heavier 3d meshes. I talked with Steve (my cohost in this stream) and we are bumping it up.

[line-weight]

8 minutes ago, Jim Wilson said:

Yes, we are tracking it as VB-145786. Originally it went in as a bug against Push/Pull and BIMObjects specifically, but on further investigation it seems more likely that it's not BIMObjects specifically, but the fact that they're normally made up of heavier 3d meshes. I talked with Steve (my cohost in this stream) and we are bumping it up.

 

Thank you, I agree it's likely affected by any complex objects rather than BIMobjects specifically. I also think that might not be the full story. But it's good to know that it's being dealt with.

[PVA - Admin]

So i've learned something I wanted to share about the On Demand Tessellation feature I demoed: When I shot that video, ODT was using only 3 additional cores to the main core, so it was topping out at using 4 CPU cores total, even though that machine had 8. So the speed improvements going forward will be even more significant than seen in the demo. I will check with engineering and ask about the current limitation and see when it is planned to increase. This will directly affect current hardware recommendations so it interests me personally, greatly.

[Mark Aceto]

11 hours ago, Jim Wilson said:

So i've learned something I wanted to share about the On Demand Tessellation feature I demoed: When I shot that video, ODT was using only 3 additional cores to the main core, so it was topping out at using 4 CPU cores total, even though that machine had 8. So the speed improvements going forward will be even more significant than seen in the demo. I will check with engineering and ask about the current limitation and see when it is planned to increase. This will directly affect current hardware recommendations so it interests me personally, greatly.

 

This is exactly my point when comparing a 2017 iMac with a 2017 iMac Pro.

 

Would also like to understand the performance usage/difference between an 8GB and 16GB GPU (not sure how to measure that).

 

As of August 2018, it seems like spending over 5k on any Mac won't yield much return, so September will be very interesting... 

[zoomer]

4 hours ago, Mark Aceto said:

Would also like to understand the performance usage/difference between an 8GB and 16GB GPU

 

It is basically the same as with RAM - you need to have "enough" of it.

If you have too less of RAM it needs to use the disk to swap, which is far slower.

If you want to display a model by a GPU, the whole model data has to fit in.

If you want to render on a GPU and the whole model does not fit in, it won't eben

work at all.

So if you already have "enough" RAM or VRAM, adding more to it will not accelerate

in any way.

It is just that you want to have enough memory for the whole lifetime of your machine

where the demand is likely to increase over time. Especially as it is harder to upgrade

later than it was in the past as there is a trend to seal and solder hardware.

[zeno]

Great improvement! Thank you! It will increase also the pointcloud mangment? They depend of RAM or GPU? 

[zoomer]

Basically both RAM and VRAM.

But I think handling and file size of Point Clouds is quite efficient (?)

and will not be as demanding as one would think.

Too less VRAM should be very slow but you could reduce density

when importing or crop your cloud later if needed.