Houdini 22.0 Reference Windows

Copernicus Default Settings window

This window lets you change the global context options for Copernicus, which sets the default settings of the COP nodes. These control the default settings of COP nodes that don’t have their own input or values for these settings. You can also override some of these settings by turning on the related parameter in the COP Network and COP Network SOP. For more information about the default parameters, see Default COP Network settings.

Note

The COP nodes default to the network settings in the following order:

  1. Inputs and values for these settings on the individual node

  2. COP Network and COP Network SOP

  3. Copernicus Default Settings window

These settings only affect the current scene file. To make the current scene file’s settings the defaults for all future Copernicus projects, click Save as Default.

For information about the COP network, see Copernicus.

Parameters

Resolution

The global context option for default_xres and default_yres. This is the resolution size of a layer.

Precision

The global context option for default_precision. This is how many bytes of storage each channel of a layer requires. More bytes allow for larger integer values or more accuracy for floating point values, but require more memory.

16

The channel uses 16-bits and defaults to half-precision.

32

The channel uses 32-bits and defaults to full-precision.

Border Mode

The global context option for default_border. This is the border type of a layer.

See Border types for more information.

Constant

Values outside of the layer evaluate to 0.

Clamp

Values outside of the layer evaluate to the closest pixel in the layer.

Mirror

Values outside of the layer reflect off of the boundaries.

Wrap

Values outside of the layer wrap around the boundary.

Pixel Scale

The global context option for default_pixelscale. This is a scale for the size of the pixels in a layer. This value causes the image to cook at a fractional effective resolution without changing any pixel-relative values.

Video Ram

The global context option for % VRAM. This is the percentage of your current video RAM to use to store a layer’s pixels and data. Copernicus starts moving layers and VDBs from the GPU to the CPU when this limit is reached. Set a low percentage to aggressively free up GPU memory. If you're using Copernicus with other GPU-heavy applications or a background process XPU for example, you may want to set this to 10%.

Note

Shared memory or CPU OpenCL modes should use 100% because there’s no benefit to moving off of the GPU.

Low numbers (such as 10%) improve Copernicus' interaction with other applications, but can decrease the speed of large Pyro simulations. High numbers (such as 100%) make Copernicus keep items on the GPU for as long as possible, removing them only when out-of-memory events occur within the same process. High numbers provide the optimal performance.

Tip

Turn Draw Time on in the viewport to show OpenCL swap events as a yellow notice above the FPS. This lets you see if Copernicus is being overaggressive in transferring and has to copy back and forth, which degrades performance.

File Cache

The amount of megabytes to devote to File COP caches, which cache loaded AOVs in CPU RAM to avoid slow file formats or disk reads.

Auto-Unload

This feature is still under development. The current functionality is unfinished and subject to change, and may have thin or no documentation. Please bear this in mind when using it.

Discards node outputs during or after the cook to minimize RAM usage. To prevent recooking reused nodes, this uses various heuristics to keep nodes cached.

This setting is useful because, by default, Copernicus caches every node’s output data. This default behavior lets you move around the network without recooking, but uses considerable CPU RAM. This also uses a lot of Video RAM because the data is kept on the GPU until the Video Ram limit is reached.

See the copcache -U HScript command for more technical details.

Never

Cache all nodes and don’t attempt to unload them.

After Cooking

Unload all used nodes once a top-level cook event is complete.

Used Inputs

As each node completes cooking, unload the node’s input if the input has all of its outputs used.

Always

Unload inputs and do a final unload after the top-level cook.

Windows