Just saw the following, looks like just a mem over-allocation attempt. Repop is on (subdivs =3, p/sub = 4), particle population is about 2 million.
…so this was a big attempted memory allocation of 122GB, about double what was free at the time. System in question is HP z820, 96GB, K5000. First such crash with this project.
Posting FYI; I don’t really consider memory issues to be “bugs” so much as just a act of life.
Definitely the issue here is that the repopulate “Fill-radius” is set small in combination with a high “Subdivisions” value. This resulted in producing far too many particles for the system to handle.
A good rule of thumb is. Prior to rendering:
a) Set the “fill radius” to something quite large… For example, if your particles approximately occupy a 100 unit sphere, maybe start with a fill radius of 1.
b) Always start with a “subdivisions” value of 0.
c) Start with a “Number of particles per subdivision” value of 1.
Render, then do the following:
a) Slowly reduce the “fill radius” so the result doesn’t look like ping-pong balls.
b) Then increase the “subdivisions” by 1 at a time until the volume starts to look more solid. Increasing this number exponentially increases the particle count.
c) Lastly, double the “Number of particles per subdivision” until you have a really nice looking smooth render.
I hope that helps! The problem here is that it is really hard for us to provide good default values, and if the user just starts entering numbers, it is quite common to use up all the computer’s memory. We are working on a way to provide better defaults and some viewport previews to make everyone’s life easier.
Thanks for the report!