
One of the first lessons I ever had with nuke is how unstable it can be. It allows you to bog down your system, it allows you to create things that are orders of magnitude more computationally expensive than you need, and it almost encourages you to run betas and bleeding edge releases that have a fair amount of bugs in them. It’s funny to admit, but that isn’t always a bad thing either! Being able to push it to the limit allows us to push our work further as well. That’s the price we pay for having cutting edge tools. Sometimes instability is a fact of life.
One thing that I absolutely love is a little docklet app for osx called istat menus. It allows me to look at my ram usage (including swap) cpu usage, network transfer, and hard drive free space. All of those can come into play when you have an unstable system. My most frequent offense is having too many instances of nuke open and filling up the ram with cache files. Eventually that leads to utilizing swap, which is when it starts writing ram to the hard disk in order to make more space. That makes everything very, very slow. With istat I can quickly glance up and see if my machine is slow because I am simply hammering the processor, or if I am bottlenecked by swap. If it’s swap, then the answer is simple: close as many instances as possible to release their cache files. If it’s the processor then it is more complicated, but at least I know where to look.
I hear that the linux version of nuke has a built in tool for diving into which node is using the most resources, but I’ve never worked on linux. I would really love to see that in action.