Boxer: parallel, scalable - everywhere

Boxer 2.0 is parallel everywhere

• parallel file I/O
• parallel octree meshing
• parallel topology discovery and geometry editing
• parallel level-set solver
• parallel hybrid-mesh generation from the core octree
• parallel mesh quality control during body-conformal export
• No serial bottlenecks

Not just multi-threaded, but distributed

• benefits from machines with more than one CPU/core - but not limited to a single machine
• can access shared and/or distributed memory resource simultaneously

Parallel and distributed means scalable:

• can run on anything from 1 to 10,000+ nodes and/or CPUs
• the licensing model of CFS is not simply per-node as with other vendors - users can run large parallel jobs without paying extra for every single new CPU deployed