OpenVDB Introduces SIMD Framework With Some 2~4x Performance Improvements

OpenVDB is the sparse volume data structure and tooling library maintained by the Academy Software Foundation. OpenVDB in turn is used by various CGI software for dealing with sparse volumetric data such as Houdini, RenderMan, and Cinema 4D to the open-source Blender. It's even won an Academy Award for technical achievement while now in 2026 it's finally establishing a SIMD framework for better leveraging modern x86 ISA capabilities. OpenVDB 13.1 was released today by the Academy Software Foundation and one of the main changes is the introduction of a SIMD framework for targeted ISA vectorization support. This is for explicit intrinsic support on x86 and initially is making use of Agner Fog's VectorClass Library with ISA capabilities up to AVX-512. When OpenVDB is built with the new USE_VCL=ON option for utilizing the VectorClass Library, it's yielding some mighty performance wins for supported operations. Initially the point transfer algorithms have been adapted to use SIMD instructions and is yielding from 2x to 4x performance improvements when running on modern x86_64 CPUs equipped with AVX-512. More details on OpenVDB's new SIMD support can be found via this pull request. Hopefully future releases will continue building out the SIMD support further to cover more operations. OpenVDB 13.1 is released on GitHub with this initial SIMD frameworrk. The new release also brings NanoVDB enhancements around CUDA resource management, faster kernels, and more. The release is rounded out by improvements to the vdb_tool command line program.