High Performance Computing

Need the highest performance to overcome runtimes associated with large simulations? FLOW-3D CAST seamlessly transitions from its desktop version to its high performance, on-demand cloud computing and cluster version to cater to the industry’s most demanding simulation needs. If you are interested in learning more about the HPC version of FLOW-3D CAST, please contact us at hpc@flow3d.com to schedule a one-on-one demo.

High Performance Computing Benchmarks

HPC benchmarks for FLOW-3D CAST v5.0 up to 600 cores for typical casting applications on FLOW-3D CLOUD using the latest Intel® Skylake processors.

High Pressure Die Casting Suite

High Pressure Die Casting Suite
High Pressure Die Casting Suite FLOW-3D CAST HPC benchmark

In this simulation, a high-pressure die casting case was studied and performance of FLOW-3D CAST was reported for the filling process. A 24 hour simulation on 40 cores was reduced to 10 hours on 640 cores.

Geometry features: Steel mold, copper plunger, overflows, covers, vents and cooling channels
Mesh: 6.6 million cells
Physical Models: Gravity, heat transfer, surface defect tracking, air entrainment, cavitation, viscous flow with RNG turbulence model, and solidification
Pressure Solver: GMRES

Permanent Mold Casting Suite

Permanent Mold Suite
Permanent mold suite FLOW-3D CAST HPC benchmark

In this simulation, a permanent steel mold casting was studied and performance of FLOW-3D CAST was reported for the filling process. A 24 hour simulation on 40 cores was reduced to 9.25 hours on 640 cores.

Geometry features: Carbon-steel inserts, insulating risers and sleeves, cooling channels and steel mold
Mesh: 2.7 million cells
Physical Models: Gravity, heat transfer, surface tension, surface defect tracking, air entrainment, viscous flow with RNG turbulence model, and solidification
Pressure Solver: GMRES

Sand Casting Suite – I

Sand Casting Suite
Sand casting suite HPC benchmark

In this simulation, a gravity sand casting process was studied and performance of FLOW-3D CAST was reported for the filling process. A 24 hour simulation on 40 cores was reduced to 12.7 hours on 320 cores.

Geometry features: Sand mold, filters, insulating sleeves and air caps
Mesh: 9.8 million cells
Physical Models: Gravity, heat transfer, surface tension, surface defect tracking, air entrainment, viscous flow with RNG turbulence model, and solidification
Pressure Solver: GMRES

Lost Foam Casting Suite

Lost Foam Suite
Lost foam casting suite FLOW-3D CAST HPC benchmark

In this simulation, a lost foam casting process was studied and performance of FLOW-3D CAST was reported for the filling process. A 24 hour simulation on 40 cores was reduced to 6.6 hours on 640 cores.

Geometry features: Lost foam, sand mold
Mesh: 4.3 million cells
Physical Models: Gravity, heat transfer, surface tension, surface defect tracking, air entrainment, viscous laminar flow, solidification, full lost foam model with defect tracking
Pressure Solver: GMRES

Sand Casting Suite – II

Engine block sand casting FLOW-3D CAST
Sand casting suite HPC benchmark engine block

In this simulation, a sand casting process was studied and performance of FLOW-3D CAST was reported for the gravity filling process of an engine block. A 24 hour simulation on 40 cores was reduced to 6.5 hours on 640 cores.

Geometry features: Sand mold
Mesh: 4.3 million cells
Physical Models: Gravity, heat transfer, surface tension, surface defect tracking, air entrainment, viscous laminar flow, solidification, full lost foam model with defect tracking
Pressure Solver: GMRES