Resample the particles using the parameters below in order to compensate for irregularities in distribution of particles. This can reduce the number of particles in over-sampled areas, giving better performance and preventing artifacts.

The particle separation used in the simulation. This is used as a unit length for the following next three parameters.

The voxel side length to use for the generated VDB volume. This is a scale on the Particle separation length. For example, if Particle separation is 0.1 and Voxel size is 0.5, the side length of the voxels of the output field will be 0.05.

The maximum distance at which particles interact. Small increases can give smoother results but increase the cooking time and memory usage greatly. This is a scale on the Particle separation length. For example, if Particle separation is 0.1 and Influence radius is 3, particles will interact if they are within 0.6 units of each other.

The approximate desired distance between the particles and the generated surface. This is a scale on the Particle separation length, and must be smaller than the Influence scale. It becomes a scale on the pscale attribute value (instead of Particle separation length) if pscale is present.

You can use the pscale attribute to locally manipulate the distance of the surface to particles (pulling it closer or pushing it away). The pscale value replaces Particle Separation as the scale by which this parameter is multiplied to calculate the indented distance of each particle from the surface of the fluid.

You must ensure the Droplet Scale times pscale value of all particles is smaller than Influence Radius times Particle Separation.

The size of the cubic “tiles” used by the algorithm. These tiles are visualized in the viewport when the display flag on this node is on. Tiles can be processed in parallel, so if you have many CPUs you may want smaller tiles. There is however an overhead in distributing the work into tiles which increases with having more tiles. The amount of memory used by the node is related to tile size and not to the number of tiles. Larger tiles increase the amount of memory used by the node.

This value is a scale on the Influence Radius parameter on the Particles tab. So, if you keep the Influence Radius the same, doubling this value increases the memory usage 8 times. Similarly, halving the voxel size would also increase memory usage 8 times.

The displacement of the origin of the tile array from the grid origin. Changing this parameter shifts the entire array of tiles by the specified amount (rounded to the nearest multiples of voxel size). Modifying this parameter can reduce the number of tiles which overlap the main body of the fluid and this can noticeably improve the performance of the node. For example, if the fluid simulation contains a shallow pool of water, it is a good idea to have this main pool (ignoring the splashes) covered by as few rows of tiles as possible.

The amount of extra overlap between adjacent tiles. The larger the padding, the lower the risk of inconsistencies between tiles, but at the same time, the more the amount of work the node has to do per tile. Increasing this value can greatly increase the cooking time of the node. Increasing this beyond 3 generally has little to no benefit.

This value is a scale on the Influence radius parameter on the Particles tab.

If turned on, the maximum percentage of the computer’s total memory to use for the node’s work buffer.

This does not affect the amount of memory used by the field created by the node.

Large Tile size and Additional padding values may make it impossible to fit within this restriction. In that case, the node will stop and display an error. If the amount of memory is too small to use all available CPUs, the node will display a warning.

The choice of visualization of Tiles or Padded Tiles (or both) used for the breakdown of the surfacing job.