I have asked for this as well. If this could be done it would be a tremendous help. Layout is only using one core as it stands right now so use another core for text while the other one renders.
In my understanding that is not correct, but again I could be wrong. -n is needed to force open to start a new instance.
I don’t think there is a simple way to tell which core an app is using, and the OS may actually shift an app from core to core depending on load. But if you look at the Activity Monitor and see two instances of SketchUp, they are separate sessions and could potentially be spread to separate cores.
I see a nice row of LayOut icons in my Dock. That makes me think they are different instances.
Hmm. Maybe using the path to the executable inside the app bundle bypasses the normal behavior.
I think that @slbaumgartner 's comments on performance and running multiple concurrent instances of Layout would affect the aggregate performance of all the instances, but not the performance of one instance. For example, if a certain Layout task takes X minutes to do when there is only one instance of Layout running, then that same task would still take about X elapsed minutes even when there are additional instances of Layout running.
On the other hand, if you ran three instances of Layout and executed that same task in each instance concurrently, the elapsed time to completion would probably be about X minutes (assuming no significant sharing of resources between instances), rather than the X*3 minutes elapsed time that would apparently be required if the three tasks were performed consecutively within one instance of Layout.
Said another way, running multiple application instances on a system with multiple CPU cores allows each instance to run concurrently. That can decrease elapsed time for aggregate workloads spread across the instances, but it does not speed up any one instance.
I have been struggling with this technical problem for years.
I’m even thinking of switching to another software…
I don’t think it will be the best CAD software until the problems with Layout software are improved.
What I would like to see in Layout is
- Faster vector display
- Improved compatibility with other CAD data, e.g. vwx (text and dimension reading)
- Improved graphic description (pathfinder-like functions such as integration)
- Display of arc dimensions
I believe that even if these improvements are made, user acquisition will increase.
I sincerely hope that the software will become more professional.
Nice analysis and suggestion list AK_SAM, I hope @Adam is tracking this thread! One can only hope for some progress soon!
Yeaaah this stuff’s been discussed infinity times, for years.
The topic is apparently locked on with Trimble’s short range radar, sonar, and the infrared deep space scanner. Staff are tracking it on GPS, have bluetooth enabled, and they employ an expert indigenous tracker.
Dog the Bounty Hunter is also on the case.
So yeah, awareness is high.
Each year as the new release date approaches, there are two questions always on my mind:
- What market (user base) is SU+LayOut catering to? (makers, concept designers, builders…maybe these groups now have diverging priorities).
- How much of Trimble’s vast resources is it devoting to the issue?
We will see…2022 may be an important year for SketchUp Pro and LayOut.
I am on an 8mb m1 mac. It was the only machine I could get when my previous mac died. I would say I was in the Layout is slow camp, until I really had to rely on it day in and day out and that was on 2015 imac. With some use, I learned its quirks and features. Now on my m1 base, I find that it is not slow at all.
However, for much of what I do now, I am exporting ISO images to Affinity Publisher for some killer presentations.
Again, layout is not slow, it just takes a bit to get used to a workflow and then it is lovely.
Would love to hear about your Affinity workflow?!
I have the same frustrations and experience from running SU+LO on both M1 and M1 Pro macs (and my previous intel-macs too). It totally feels like the ball is rolling backwards, and even though I get better and better to build optimalized models with few textures and rational geometry, it still never becomes fluid to work in Layout.
As you say, raster is not sufficient, and vector shouldn’t be that difficult to use and horribly slow.
We’ll just have to continue to pay and hope that something good is in SU’s pipeline, or if some good alternative emerges.
Why does a surface modeler have this limitation? And, what is that compared to?
Is not Blender a surfacing modeling software? Pardon my ignorance, just want to better understand why this always comes up.
this is something that I would push for. not sure how it would work, but being able to set a scene to export as flattened would be a great feature for the 2D work in a set. This would also address the other point about snapping issues with invisible edges.
The terminology here is somewhat inaccurate. SketchUp is a face modeller and what are called surfaces in in are collections of multiple faces with the edges between them softened. Faces and their defining edges in SketchUp are static.
A Surface modelling application (like Rhinoceros or Inventor or…) works with mathematically defined curves, using them to generate surface patches in all possible shapes, using generated faces just to display the results, with a density that can vary.
Usually Solid modellers (like Solidworks) use a similar way to generate shapes, but deal only with closed surfaces that then are called solids.
SketchUp is actually more specific than just a “surface modeler”. It is actually a polygonal surface modeler. That means that everything is built using flat, polygonal faces bounded by straight edges and that when a closed solid object is formed, there is no material inside it. It uses visual tricks to give the appearance of smooth surfaces, but they are actually faceted. Some other surface modelers are based on NURBS representations, which are smooth objects in themselves.
Edit: I see @Anssi wrote much the same while I was typing,
so are there any true alternatives to fix the bezier/circle issue within Sketchup? Seems like that comes up a lot but if it’s a foundational function of the software it can’t change unless the software is re-coded from the foundation. Am I wrong or just simple minded?
I guess my other question is what’s the tradeoff when modeling in the other softwares? I know I have tried Blender, Inventor, and Fusion 360 but found them to be more complex in usage. Is that the main difference? Do they have similar issues getting models into 2D drawing sets?
You are correct: the internal representation of everything in SketchUp is built on straight lines and flat polygons. A total rework would be required to provide anything else.
I don’t necessarily recommend this process, but a version of this does currently exist. Select a viewport and right click > Explode. This turns the linked Sketchup Viewport into a scaled drawing on the paper space. You can still pull accurate dimensions from the drawing, you can still drag the drawing section around like a viewport, even still change the scale (size on paper) via the Scaled Drawing Window, and by double clicking on a drawing you can draw on top of it in scale for adding or making notes. It really retains many of the layout properties of a viewport. Just as you say, by exploding a viewport you loose all connection with the 3D SketchUp model info so you can no longer change views or adjust style, tag visibility, line type, hybrid render, etc.
As a test I took a 53MB Layout file with about 12 viewports in it, exploded them all, went to File>Document Setup> References > Purge (if a reference won’t purge then it has an unexploded viewport somewhere). The resulting Layout file looks the same and is 727kb.
Again, I’m not advocating this option, there many reasons why this is a compromising process but it accomplishes some of your request so I mention it.
interesting…That’s a dramatic decrease in size.
breaking the reference is obviously not great.
So it has been said. However, Rhino does support polygonal, surface and quad modelling in a single, OpenGL-based application.
I think it may be easier if from the start you plan a different geometry engine than SketchUp did. The various types supported by Rhino are all expressible as mathematical models so I would expect they can all be handled as specific cases of a general representation. But SketchUp wasn’t designed that way. It was committed to simpler math right from the start.