Running at 65% of Peak.

This software performs Dissipative Particle Dynamics simulations of blood and cancer cell separation in complex microfluidic channels with subcellular resolution. It relies on CUDA and MPI to run efficiently on GPU-accelerated supercomputers, for example performing at 65.5% of the available 39.4 PetaInstructions/s in the 18'688 nodes of the Titan supercomputer.

Outperforming the competition by 45X.

For shear layer flow benchmarks, this software constantly outperforms the GPU package of LAMMPS by 30-45X, varying the number of nodes from 27 to 18'600. In the limited range of nodes supported by HOOMD-Blue (up to 280 for our benchmark), we report a constant outperforming factor of 4X.

Devising the next generation Lab-On-A-Chip.

The simulations delivered by this software outperform by one to three orders of magnitude the current state-of-the-art simulations in terms of numbers of red blood cells and computational elements. The software emulates the conditions and the geometric complexity of microfluidic experiments. These simulations provide sub-micron resolution while accessing time scales relevant to engineering designs.

Simulation of the CTC-iChip1 uDeviceX movie. Visualization by Christian Conti.

This is a Free software.

The software comes with the GNU GPL v2.0 license. This forces you to share your improvements too.

This work, including the development of the software, started at the CSE Laboratory of ETH Zurich, headed by Prof. Petros Koumoutsakos. The team included researchers from ETH Zurich, Brown University, Universita da Svizzera Italiana (USI) and Consiglio Nazionale delle Ricerche (CNR).

Related publications:

  • Rossinelli, D., Tang, Y.-H., Lykov, K., Alexeev, D., Bernaschi, M., Hadjidoukas, P., Bisson, M., Joubert, W., Conti, C., Karniadakis G., Fatica, M., Pivkin, I., Koumoutsakos, P., The In-Silico Lab-on-a-Chip: Petascale and High-Throughput Simulations of Microfluidics at Cell Resolution, ACM 2015 Gordon Bell Award Finalist, Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis (SC '15), 2015 (doi) (pdf)
  • Contributors: Alexeev, D.; Amoudruz, L.; Bernaschi, M.; Bisson, M.; Conti, C.; Economides, A.; Fatica, M.; Hadjidoukas, P.; Joubert, W.; Karniadakis G.; Koumoutsakos, P.; Kulakova, L.; Litvinov, S.; Lykov, K.; Pivkin, I.; Rossinelli, D.; Tang, Y.-H.

    The contributions of the individual team members can be found here.