GPU-based flow simulation with detailed chemical kinetics

A methodology for the integration of stiff chemical kinetics on GPUs

“…Finally, we point out that the above approach is at variance with the shared memory (reduced storage pattern) scheme tested by Le et al [12]. In the context of a one-block approach, these authors suggested to store the pivot row and the row which is currently being updated in CUDA shared memory.…”

“…(1) in the above, the reaction fractional step in Eq. (12) is not particular to this model, but can be constructed similarly for any scalar or vectorvalued diffusion-convection-reaction transport problem. Specifically, the OpenCL-GPU solver for the reaction fractional step which we present is applicable to turbulent flow problems and can be readily incorporated into an existing RANS or LES reactive flow solver.…”

“…For the one-thread approach, this implies that one work group requires a local memory buffer of size ln c which may be incompatible with the maximum allocatable local memory buffer size or may impact the compute unit's occupancy. Le et al [12] discussed this issue in more detail and quantified limits on the reaction mechanism size for the case where the array that one work item stores in local memory is of the order of n c or larger. On the plus side, implementing the one-thread approach does not require any change in the serial implementation of the integration scheme or the chemical kinetics source term evaluation.…”

“…Le et al [12], for instance, included a first order accurate backward Euler scheme in a CUDA-based high order finite volume solver for the reactive Euler equations. As Table 1 GPU implementations of explicit/implicit integration schemes (double precision) for chemical kinetics ODE systems.…”

GQSD: The program for the graphic processing units accelerated quantum scattering dynamics

“…Two approaches can be followed: new mathematical methods and parallel computing on latest hardware platforms. The GPU has hundreds or thousands arithmetic elements to perform massively parallel calculation, and has been used in many computational areas of theoretical chemistry, such as molecular dynamics, [5][6][7] quantum chemistry, [8][9][10] chemical kinetics, [11,12] and potential energy surface. [13] Recently, GPU has also been applied to quantum scattering calculations using both TID and timedependent wavepacket methods.…”