Added support for a file-based Tractor scheduler login. $PDG_TRACTOR_PASSWORD_FILE can be set to the path to a text file containing the base64-encoded password.

Fix hanging issue that could occur when creating a node with a name that already exists and its name has a numeric suffix that exceeds the signed 32-bit limit (2,147,483,647).

Fix hanging issue that could occur when creating a node with a name that already exists and its name has a numeric suffix that exceeds the signed 32-bit limit (2,147,483,647).

Fix hanging issue that could occur when creating a node with a name that already exists and its name has a numeric suffix that exceeds the signed 32-bit limit (2,147,483,647).

Fix hanging issue that could occur when creating a node with a name that already exists and its name has a numeric suffix that exceeds the signed 32-bit limit (2,147,483,647).

Fix hanging issue that could occur when creating a node with a name that already exists and its name has a numeric suffix that exceeds the signed 32-bit limit (2,147,483,647).

Fixed a bug that prevented the Convert Tets SOP from working correctly for 64-bit precision positions. Additionally, Convert Tets preserves the original relative order from the points of the incoming tet mesh.

Fixed a bug that prevented the Convert Tets SOP from working correctly for 64-bit precision positions. Additionally, Convert Tets preserves the original relative order from the points of the incoming tet mesh.

Fixed a bug that prevented the Convert Tets SOP from working correctly for 64-bit precision positions. Additionally, Convert Tets preserves the original relative order from the points of the incoming tet mesh.

Fixed a bug that prevented the Convert Tets SOP from working correctly for 64-bit precision positions. Additionally, Convert Tets preserves the original relative order from the points of the incoming tet mesh.

The primuv() VEX function will now properly interpolate attributes in 64-bit mode. In particular, this affected lense shaders as they run in 64-bit mode by default.

SOP Import now supports importing 16-bit and 64-bit float attributes where the type info is point, vector, color, normal, or quaternion. Previously, these were always imported as 32-bit attributes.

SOP Import now supports importing 16-bit and 64-bit float attributes where the type info is point, vector, color, normal, or quaternion. Previously, these were always imported as 32-bit attributes.

SOP Import now supports importing 16-bit and 64-bit float attributes where the type info is point, vector, color, normal, or quaternion. Previously, these were always imported as 32-bit attributes.

SOP Import now supports importing 16-bit and 64-bit float attributes where the type info is point, vector, color, normal, or quaternion. Previously, these were always imported as 32-bit attributes.

SOP Import now supports importing 16-bit and 64-bit float attributes where the type info is point, vector, color, normal, or quaternion. Previously, these were always imported as 32-bit attributes.

When using OpenSubdivision with 64-bit floating point positions, the results will now be preserved in 64-bit floating point precision.

When using OpenSubdivision with 64-bit floating point positions, the results will now be preserved in 64-bit floating point precision.

When using OpenSubdivision with 64-bit floating point positions, the results will now be preserved in 64-bit floating point precision.

When using OpenSubdivision with 64-bit floating point positions, the results will now be preserved in 64-bit floating point precision.

Added support for primvar AOV to karma (64-bit primvars are downcast to 32-bit). VEX export AOVs no longer require presence of an LPE AOV to work.

The OpenGL ROP uses 64b precision to compute the view transform so its frustum matches that of the viewport.

The OpenGL ROP uses 64b precision to compute the view transform so its frustum matches that of the viewport.

The OpenGL ROP uses 64b precision to compute the view transform so its frustum matches that of the viewport.

The OpenGL ROP uses 64b precision to compute the view transform so its frustum matches that of the viewport.