One of my artists have a hip file that weight over 2gb… There's a lot of nodes in there. I'm trying to find what is weighting so much. Did he lock or stash a node? Is there a tool I can use to analyze the scene and find where the heavy data resides in his scene?

You can try hexpand [www.sidefx.com] to expand the hipfile into a regular file structure on disc and then use normal shell tools to sort by size.

I've found hexpand in the HFS, yet I have no idea how to actually use that. I'm on windows, and the various standalone .exes I can use, but hexpand is one of those without a file extension. Are those tools limited to linux only?

Thanks in advance,
Martin

Before you go to hexpand, yes, look for stash nodes. A quick way of doing this is open a textport (Alt Shift t) and type


This will show you where all the stash nodes are.

You can also lock nodes, which are not as easy to find. To track these down, again in the textport:


… then in the directory where you launched Houdini, there will be a text file that describes your scene. Now do a simple find for “-l on” which is “hard lock”. These nodes can be massive too.

If you haven't found anything, you can again, through the textport, rebuild the scene from scratch as if you laid down every node individually. Just open the textport in a blank houdini, and source the debug.txt file you created earlier. This is a quick and easy way to rebuild corrupt files.


Still no joy? You could try hexpand through (you guessed it), the texport (cuz I have no idea about Windoze):


… although it does not seem to work on .hipnc files which is what I have to test on.

Sean

Hello Sean,

thank you a lot for this detailed explanation! Haha, using system() in the textport would have taken me an eternity to figure out. It's a real pity that it doesn't work on NC files, as we're also in a university. But the opscript way will have to do.

Thanks again,
Martin