Ruby efficiency for large number of operations

tt_su · 18 Apr 2014, 16:01

If I make each face be its own group (i.e. make a polymesh INSIDE the loop and fill_from_mesh inside the loop as well) it goes slooooooow. Ideally for my plugin, each one of these would be their own group (of course, what I originally was trying to do is a bit more complicated so each one would be more than a single plate).. I wonder if there's a way to get each one to be a separate group but still be reasonably fast?

Got a sample snippet for this? When optimizing for speed it's hard to talk generically. Having a common set of sample code to refer to will help.

@drewmorgan said:

Based on our previous discussion, I'm guessing once you "fill_from_mesh", it then checks the points that are read in later will all the previous ones, whereas if you just put them all in the polymesh it doesn't do those checks?

I'm sorry, but I don't quite follow what you mean here.
fill_from_mesh require that you add the PolygonMesh to an empty Entities collection. It then simply builds the PolygonMesh without doing any more magic.

@drewmorgan said:

If there's more than one implementation of the ruby function in the same directory, SKP will not provide an error message, but will pick one and use that.. This made me think I was losing my mind as I was modifying stuff left and right and not seeing any updates. Turns out when I created the file to post here, I just wrote a subset of my main plugins file to the same dir. All my mods were to my main file, but apparently, SKP was using the (unmodified) implementation in the subset file.. Argh. That took me far longer to figure out than it should have.

That's Ruby for you, it let you override anything. The last one to load will be the one running.

drewmorgan · 18 Apr 2014, 16:25

@tt_su said:

@drewmorgan said:

The program you provided is definitely a lot faster and seems to work well, except for in the 20000 face test case, it causes SkecthUp to crash.. It seems to be during the fill_from_mesh call, because it prints out a message to console right before that happens.. But then I get a window that pops up saying "SketchUp Application has stopped working", and it crashes.. It seems to work on some test files I have with a lower number of faces, so maybe its just too many faces for a single polymesh? Does that seem reasonable, or do you think the crash is due to something else?

20000 faces shouldn't crash SU, no.
Did you submit the BugSplat? If you did, did you enter some details that I can use to look it up?

When I get the crash, I don't get an option to send in the BugSplat report. Do I have to activate that somewhere? I poked around in the preferences, and maybe missed it? I'm only given the option to "Close the program" or "Close the program and look for a solution".. Down in the "details" section, it mentions BugSplat.dll but there's no option to submit it.. Sorry to be dense!

drewmorgan · 18 Apr 2014, 16:40

@tt_su said:

@drewmorgan said:

If I make each face be its own group (i.e. make a polymesh INSIDE the loop and fill_from_mesh inside the loop as well) it goes slooooooow. Ideally for my plugin, each one of these would be their own group (of course, what I originally was trying to do is a bit more complicated so each one would be more than a single plate).. I wonder if there's a way to get each one to be a separate group but still be reasonably fast?

Got a sample snippet for this? When optimizing for speed it's hard to talk generically. Having a common set of sample code to refer to will help.

@drewmorgan said:

Based on our previous discussion, I'm guessing once you "fill_from_mesh", it then checks the points that are read in later will all the previous ones, whereas if you just put them all in the polymesh it doesn't do those checks?

I'm sorry, but I don't quite follow what you mean here.
fill_from_mesh require that you add the PolygonMesh to an empty Entities collection. It then simply builds the PolygonMesh without doing any more magic.

Sorry - not very clear.. What I meant was, in fredo6's posted script, he creates one big PolygonMesh outside the loop over faces, and calls "add_polygon" for each face inside the loop.. Then, after the loop, he creates a new group and calls fill_from_mesh to include all the faces all at once.. That seems to be working fairly quickly. However, if I move the creation of the PolygonMesh inside the loop, add a single face to it, create a new group, and call fill_from_mesh all inside the loop, it crawls and eventually hangs like it was originally..

So, this seems quick:


    model = Sketchup.active_model

    numItemsLine = testFile.gets
    numItems = numItemsLine.split(" ")[2].to_i

    polyMesh = Geom;;PolygonMesh.new(4 * numItems, numItems)
    for i in 0..numItems - 1
      ##...
      polyMesh.add_polygon [face0Vert0, face0Vert1, face0Vert2, face0Vert3] 
    end

    newGroup = model.active_entities.add_group
    newGroup.entities.fill_from_mesh polyMesh, true

but this is slow:


    model = Sketchup.active_model

    numItemsLine = testFile.gets
    numItems = numItemsLine.split(" ")[2].to_i

    for i in 0..numItems - 1
      ##...
      polyMesh = Geom;;PolygonMesh.new(4, 1)
      polyMesh.add_polygon [face0Vert0, face0Vert1, face0Vert2, face0Vert3] 
      newGroup = model.active_entities.add_group
      newGroup.name = "itemNum" + i.to_s
      newGroup.entities.fill_from_mesh polyMesh, true
    end

tt_su · 18 Apr 2014, 16:41

Oh! You're on Windows and you don't get the BugSplat window? :s
Can you send me the script that cause that?

tt_su · 18 Apr 2014, 16:43

@drewmorgan said:

not very clear.. What I meant was, in fredo6's posted script, he creates one big PolygonMesh outside the loop over faces, and calls "add_polygon" for each face inside the loop.. Then, after the loop, he creates a new group and calls fill_from_mesh to include all the faces all at once.. That seems to be working fairly quickly. However, if I move the creation of the PolygonMesh inside the loop, add a single face to it, create a new group, and call fill_from_mesh all inside the loop, it crawls and eventually hangs like it was originally..

I see. I need to look into that.
Have you tried this with using start_operation and using the disable_ui argument?

drewmorgan · 18 Apr 2014, 16:51

@tt_su said:

Oh! You're on Windows and you don't get the BugSplat window? :s
Can you send me the script that cause that?

Actually, I get it with the script that Fredo posted above directly.. I get it with my own script as well, but it is essentially the same as what Fredo posted now that I've mucked with it to be the same general structure..

If I just place Fredo's .rb file in my PLugins directory then select the plugin and load the testFile.txt, I get the crash..

Here's what I get down in the "Problem Details" section of the pop-up window:


Problem signature;
  Problem Event Name;	APPCRASH
  Application Name;	SketchUp.exe
  Application Version;	8.0.11752.0
  Application Timestamp;	4ed14f2c
  Fault Module Name;	BugSplat.dll
  Fault Module Version;	3.1.0.26
  Fault Module Timestamp;	483b5de5
  Exception Code;	c0000005
  Exception Offset;	0000e0c7
  OS Version;	6.1.7601.2.1.0.256.4
  Locale ID;	1033
  Additional Information 1;	43af
  Additional Information 2;	43afc9fd33bc84f320a0c75763285cbb
  Additional Information 3;	4365
  Additional Information 4;	4365e8739e6c211b9e2e8edffa6340e4

Read our privacy statement online;
  http://go.microsoft.com/fwlink/?linkid=104288&clcid=0x0409

If the online privacy statement is not available, please read our privacy statement offline;
  C;\Windows\system32\en-US\erofflps.txt

and here's what I got in the console before the crash occurred:


Reading File and creating Polygon Mesh....
Nfaces = 20000 - Creating Mesh time = 6.147

Oh, I should mention this in case its related to not geting a BugSplat window.. I'm working on a corporate computer, and they are very concerned about security, so every 10 minutes or so, I get a little popup window that says something like "Can not access 0.skb".. I just dismiss it and I'm on my way, but thought I'd mention it in case its somehow related to not being prompted to submit my BugSplat.. I assume its a permissions-type thing (I can't typically write to c:\Program Files (x86), etc)..

I don't think the crash itself is directly related, because as I mentioned, Fredo's script works fine for test files with smaller number of faces, just not for the 20000 one..

tt_su · 18 Apr 2014, 19:53

Can you install and run this tool that will collect more data?

Plugins > SketchUp Diagnostics Tool > Collect Data

su_diagnostics.rbz

tt_su · 18 Apr 2014, 19:54

Oh! "Application Version: 8.0.11752.0" Are you using SU8?

drewmorgan · 18 Apr 2014, 20:04

@tt_su said:

Oh! "Application Version: 8.0.11752.0" Are you using SU8?

Yes, is that bad??

tt_su · 25 Apr 2014, 11:02

That could explain the issues you experience. Any reason you are not using the latest version? Lots of bugs has been fixed since SU8.

drewmorgan · 25 Apr 2014, 12:12

No particular reason I suppose.. I'll try to update to 14 and see if the grouping inside the loop works in the new version.. Meanwhile, thanks so much for all your help - I was able to get the single group created outside the loop to work in a matter of a minute or so for something like 60,000 faces, so I accomplished the main thing I was trying to.. For my particular issue, I'd really like each one to be a separate group so I could name it and be able to click on it and see what the name was, but I'll continue to play with it and update to 14.. Thanks again for your prompt replies and desire to help!

tt_su · 25 Apr 2014, 12:37

Glad I could help.

This particular topic of generating geometry and it's performance is a pet topic for me. I'd like to see it improved, but for now, this is the fastest way around it.

drewmorgan · 25 Apr 2014, 15:30

Yeah, your explanations helped me understand why it is the way it is.. I realize SketchUp does a lot "under the hood" in order to make it as useable as it is, so programatically adding a ton of points can be slow.. Obviously if I was manually drawing the faces in, it would likely be mostly unnoticeable. I do find it intersting though, that at a certain point it really does just hang, which makes me think even if I was interactively modeling, adding a singel face would be noticeable slow.. I can't believe that would be the case though. Luckily, I'm usually dealing with low facet count models, at least when I build them myself..

At any rate, I did download 2014 this morning.. After finally figuring out where my plugins were supposed to go (very different place than in SU8), I was able to verify that the grouping inside the loop with a call ro fill_from_mesh each time still acts the same way.. It seems to get through to about the 8000'th one (which is actually 16,000 faces for what I'm doing) and then it just stopped making progress, so I finally just killed it..

Like I said, though, the single-big-group approach will suffice for now, just isn't as ideal as the many group approach would be, so I'm happy for now.. Thanks!

tt_su · 25 Apr 2014, 15:59

hm... so even in SU2014 adding geometry in many smaller groups are slower..? This doesn't sound right...