You can simulate the fluid and temporarily store the results to the computer’s memory. This workflow can become problematic with heavy simulations, because you might run out of memory. The result will also be lost when you close the project and you have to re-simulate. To save simulation data permanently, you write them to disk. This process is called caching. Depending on factors like Particle Separation, Voxel Scale or domain size, the amount of FLIP fluid data can become very large. Houdini provides the Fluid Compress SOP node to compress the data and save disk space. Note that compression is lossy, so you will lose quality. However, SOPs such as Particle Fluid Surface SOP are designed to detect compressed fluid input and regenerate any required missing data.

1. Add a Fluid Compress node to the network. The node provides three inputs for Particles and Volumes, the Container, and Collisions. These inputs match the outputs of the FLIP Solver SOP, but in normally you probably only use the first input to compress the fluid data.

2. Connect the Fluid Compress node’s first input with the solver’s first output.

3. The Particle Separation parameter should match the FLIP Container SOP node’s equivalent.

4. If you have configured custom attribute-field pairs, append them to Keep Attributes.

Now, the data can be cached. In Houdini 19.5, all cache nodes are streamlined and provide the same base functionality and layout. In many cases you can proceed with the default settings.

1. Add a File Cache and connect it to the Fluid Compress node’s first output.

2. Base Name contains two tokens:

   • $HIPNAME is the name of the Houdini project.

   • $OS stands for the node’s own name, e.g. filecache1.

3. If you keep the entry under Base Folder, you can find the cache files in the project directory under geo.

4. To save the data, click Save to Disk (the UI will be locked) or Save to Disk in Background (the UI remains responsive).

5. To interrupt the simulation, press Cancel Cook. Depending of the simulation’s complexity, it might take a moment until cooking stops.

In simulations you should try to keep things separated as much as possible. This is also valid for caching complex simulations with fluid particles, surface meshes, and whitewater. Separated caches let you work on each part of your project individually. To do this, add multiple, cascading File Cache SOPs to your network. Then you cache the first simulation. In the next step you load the previous data from disk and perform the second simulation. Again, the File Cache SOP writes everything to disk. For the third part, you reload the data, perform the simulation, and so on. As a result all simulation data is stored separately. This way you can perform multiple surface or whitewater versions from one particle cache, for example.

Even with multiple caches, the underlying particle simulation can be compressed to save disk space. Surface and whitewater systems have built-in mechanisms to restore the original data from compressed caches.

On the lower left side of the viewport you can see two stacked icons.

1. Click the icon in the UI’s lower left corner to open the Render Flipbook settings.

2. You can leave the default settings and click Start to record an image sequence.

3. When ready, click File… in the preview window and then Save Sequence As…

4. A new dialogue appears where you can specify file name, directory, and quality.