Unser Beitrag zum Einsatz von Grafikkartenhardware im automobilen Entwurfsprozess im DIGITAL ENGINEERING Magazin ist jetzt erschienen.
Realitätsprüfung – GPU-Computing in der Automobilentwicklung
Christian F. Janßen and Thorsten Grahs, DIGITAL ENGINEERING Magazin 08-2014, Oktober/November 2014
We recently tested and implemented a grid refinement strategy in the LBM solver elbe. First, the concept was tested in the two-dimensional singlephase module elbeSP2D. The velocity field of a 2D Karman Vortex Street at Re 500 is shown below. After a proper validation of the grid coupling, the interpolation routines will be included in all the remaining elbe modules, including those for free surface and multiphase flows.
The elbe team will present the latest advances in the field of visualization and fluid-structure interaction at the 5th GACM Colloquium on Computational Mechanics (GACM 2013):
Monday, 15:20 – 15:40:
Nagrelli, Heinrich: GPGPU-accelerated simulation of wave-ship interactions using LBM and a quaternion-based motion modeler
Tuesday, 13:30 – 13:50
Koliha, Nils: Efficient Grid Generation and On Device Visualization for Massively Parallel GPU Accelerated Lattice Boltzmann CFD Simulations
Monday, 18:10, Poster Session
Koliha, Nils: Efficient Grid Generation and On Device Visualization for Massively Parallel GPU Accelerated Lattice Boltzmann CFD Simulations (Poster/Live Demo)
Moreover, this year’s symposium on Fluid Mechanics is organized by Christian Janßen and Thomas Rung (symposium #12, GACM web page). The symposium features 18 talks in four sessions on Hybrid Methods, the Lattice Boltzmann Method, Optimization, and FEM and ACM. It will take place on Tuesday and Wednesday from 10:30 to 12:10 and from 13:30 to 15:10 in building N, room 0.007 (campus map).
During my last year at the Institute for Fluiddynamics and Ship Theory at TUHH, I worked on refactoring, extending and combining my previously written GPGPU codes into one single code framework. The result is
elbe – the efficient lattice boltzmann environment
from Hamburg. It features 1D, 2D and 3D Lattice-Boltzmann kernels for various different physics, from the depth-averaged one-dimensional shallow water equations to three-dimensional multiphase models. elbe has already been used in a couple of B.Sc. and M.Sc. projects and for teaching purposes. Further information on the code, a list of publications publications and a multimedia gallery is now available at
We recently tested the new NVIDIA Kepler boards for our GPGPU flow solver elbe. After initial tests on the Kepler K20 boards were very promising and led to speed-up factors of up to 2.5x (without any code modifications), we decided to purchase the new GeForce GTX Titan board to see if the consumer hardware can keep up with the professional boards. This consumer card features 2688 streaming processors, 6 GB device memory and single- and double-precision compute capability.
Running the elbe code for both 2D/3D and single/double precision calculations, we found a performance of up to 1.5 GNUPS (giga node updates per second), 2.5x more than on our recent Tesla C2075 GPGPUs, for a D2Q9 singlephase LBM calculation.