Hi DP, I did improve the memory footprint of Grasshopper after version 0.3.0054 (the last public build). Before that, data was copied at every juncture in order to prevent components from screwing over their neighbours. I decided that the added robustness wasn't worth the additional memory overhead so now multiple components share instances of data. In some cases you'll see a marked increase in performance and decrease in memory usage because there are fewer rotting polysurface corpses lying about, but if you're actually changing the data at every juncture, it still needs to get copied obviously and there will be almost no improvement. This had to be tested rigorously, but so far the small group of internal testers have not reported any problems with this.
We'll see what happens when I unleash the new version... I've also researched some DotNET profilers recently, and maybe they can help with pinpointing the real bottlenecks. -- David Rutten Robert McNeel & Associates On Oct 1, 3:03 pm, DP <[EMAIL PROTECTED]> wrote: > That sounds great, it will most definitely help. I look forward to a > new resolved build. Do let me know if you break any grounds with the > memory issue. > > Thanks again, > > dp > > On Oct 1, 7:57 am, David Rutten <[EMAIL PROTECTED]> wrote: > > > Hi dp, > > > I solved the file reading issue. If it seems like I can't get a new > > version out before the weekend I'll send you an internal build. > > I haven't tested the memory problem yet though. > > > -- > > David Rutten > > Robert McNeel & Associates > > > On Oct 1, 1:02 pm, DP <[EMAIL PROTECTED]> wrote: > > > > Thanks David, This may also be blamed on my inefficient component > > > setup, I have noticed a few times when the memory has runout. Also, I > > > noticed that there are several "floors" at each floor, overlapping > > > many times, which could potentially multiply several times the boolean > > > operation. I am not sure how to fix this from what I suspect is the > > > BRep"Wiredframe component. This is part of the reason I posted about > > > selecting specific curves through either sorting values or strings. I > > > hope this makes sense. Thanks again for your efforts. > > > > dp > > > > On Oct 1, 4:25 am, David Rutten <[EMAIL PROTECTED]> wrote: > > > > > When opening that wrm file (without the 3dm file), I get the following > > > > errors:http://en.wiki.mcneel.com/content/upload/images/GHLoadError_TowerStud... > > > > > Obviously this is an IO bug. I don't know if I can rescue the file, > > > > through it does load fine after the errors, minus the Extrude > > > > components. So the definition becomes crippled, but it sounds like all > > > > you have to do is rehook 5 components to fix it. I'll see what I can > > > > do. > > > > > I did notice you have dozens upon dozens of BooleanDifference > > > > components strung together. This is going to take a long time to > > > > compute no matter what. It might also cause it to run out of memory. > > > > Once I fixed (or once I've given up on) the load error, I'll have a > > > > deeper look at that. > > > > > -- > > > > David Rutten > > > > Robert McNeel & Associates > > > > > On Oct 1, 11:16 am, David Rutten <[EMAIL PROTECTED]> wrote: > > > > > > DP, I'll have a look... hang on. > > > > > > -- > > > > > David Rutten > > > > > Robert McNeel & Associates > > > > > > On Oct 1, 9:32 am, DP <[EMAIL PROTECTED]> wrote: > > > > > > > I have am having trouble opening my definition file. I try opening > > > > > > up > > > > > > rhino and then loading grasshopper then when i try to load my > > > > > > definition file it takes around 5 minutes and then automatically > > > > > > closes rhino. I am hoping that I can at least recover my definition > > > > > > file...any assistance would be appreciated. > > > > > > > Please find the file > > > > > > here:http://groups.google.com/group/grasshopper3d/web/09282008_initial_tow... > > > > > > > Thanks, > > > > > > > dp
