What is SpeedIT Flow?

SpeedIT Flow™ is one of the fastest Computational Fluid Dynamics (CFD) implicit single-phase flow solver currently available. In contrary to other solutions a Semi-Implicit Method for Pressure Linked Equations (SIMPLE) and the Pressure Implicit with Operator Splitting (PISO) algorithms have been completely implemented on graphics processing unit (GPU). Based on the Finite Volume method.

SpeedIT Flow™ robust solver technology empowers users by providing accuracy in double precision on fully unstructured meshes up to 11 million cells. Our implementation was validated on standard steady and unsteady industry-relevant problems with RANS turbulence model. Its computational efficiency has been evaluated against the fastest possible CPU runs using OpenFOAM*. The results show that a server-class GPU outperforms a server-class dual-socket multi-core CPU system running essentially the same algorithm by a factor of 3.


  • Solver runs on GPU only providing high acceleration rates.
  • Supports fully unstructured meshes up to 11 million cells on a single GPU** card.
  • Accelerated pressure-velocity solution procedure for Navier-Stokes equations: SIMPLE and PISO algorithms.
  • Accurately solves steady and unsteady RANS flow cases with Model.
  • First and second order numerical schemes provide accuracy in space and time.
  • Various boundary conditions are supported.
  • AMG preconditioner and no CPU-GPU communication during calculation provide fast acceleration rates.
  • OpenFOAM(TM) reader and writer (results can be visualized with ParaView software).
  • Runs as a standalone application and is easy to configure.

GPU as an external accelerator

Now, it is possible to use your GPU as an external accelerator. You can submit your case to GPU and while it is being solved use your resources for other tasks such as pre- and post-processing.


SpeedIT Flow™ was tested against OpenFOAM software on industrial cases of steadystate and transient flow with turbulence model. SpeedIT Flow™ reduced time-to-solution up to 3.26 times compared to server-class CPU.

Results and validation