Towards blood flow in the virtual human: efficient self-coupling of HemeLB

McCullough, J. W. S.; Richardson, Robin A.; Patronis, Alex; Halver, René; Marshall, Ryan; Ruefenacht, Martin; Wylie, Brian J. N.; Odaker, Thomas; Wiedemann, Michael; Lloyd, Bryn; Neufeld, Esra; Sutmann, Godehard; Skjellum, Anthony; Kranzlmüller, Dieter; Coveney, Peter V.

doi:10.1098/rsfs.2019.0119

Cited by 18 publications

(38 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since its release in 2008, it has been applied to a range of physiological flow problems such as cerebral and retinal flow, vascular remodelling, and magnetic drug targeting [31], [32], [33]. Recent development efforts include integration of finite element and immersed boundary methods [34], [35] and efficient self-coupling [9]. While HemeLB is actively developed to take advantage of exascale The goal of the present work is to augment the existing MPI-parallelization of HemeLB with a CUDA kernel to leverage one or more GPUs available to an MPI process.…”

Section: B Related Workmentioning

confidence: 99%

“…As a result, developing a high-performance LB framework based on a hybrid CPU-GPU parallelization scheme will enable researchers to study more complex fluid flow problems in a timely manner. In this work, we describe our multi-GPU implementation of the lattice Boltzmann method, which is based on the high-performance LBM flow solver HemeLB [6], [7], [8], one of the key use cases of the MPI-4 standard [9]. We study the performance characteristics of our implementation (referred to in this work as HemeLB-GPU) with regard to several runtime parameters, and we measure the speedup and scalability of our implementation on multiple GPU architectures.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

GPU Acceleration of the HemeLB Code for Lattice Boltzmann Simulations in Sparse Complex Geometries

et al. 2021

View full text Add to dashboard Cite

We present an implementation and scaling analysis of a GPU-accelerated kernel for HemeLB, a high-performance Lattice Boltzmann code for sparse complex geometries. We describe the structure of the GPU implementation and we study the scalability of HemeLB on a GPU cluster under normal operating conditions and with real-world application cases. We investigate the effect of CUDA block size and GPU over-subscription on the single-GPU performance, and we present a strong-scaling analysis of multi-GPU parallel simulations using two different hardware models (P100 and V100) and a variety of large cerebral aneurysm geometries. We find that HemeLB-GPU achieves single-GPU speedups of 60x (P100) and 120x (V100) compared to a single CPU core, with good scalability up to 32 GPUs. We also discuss strategies to improve both the kernel performance as well as the scalability of HemeLB-GPU to a larger number of GPUs. The GPU implementation supports the LBGK collision kernel, boundary conditions for walls and inlets/outlets, and several lattice types (D3Q15, D3Q19, D3Q27), and it integrates seamlessly with the existing infrastructure in HemeLB for graph partitioning and parallelization via MPI. It is expected that the GPU implementation will enable users of the HemeLB code to make better utilization of heterogeneous high-performance computing systems for large-scale lattice Boltzmann simulations.

show abstract

Section: B Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

GPU Acceleration of the HemeLB Code for Lattice Boltzmann Simulations in Sparse Complex Geometries

et al. 2021

View full text Add to dashboard Cite

show abstract

“…In comparison, 3D models permit a far higher fidelity study of the blood flow by means of an exact representation of the domain of interest and do not rely on major assumptions about flow behaviour (e.g. symmetry) that are made by 1D solvers 10 , 11 , 16 . However, 3D simulations are more demanding to execute.…”

Section: Introductionmentioning

confidence: 99%

“…However, 3D simulations are more demanding to execute. As supercomputers become more powerful, it is more tractable to study 3D flow of large, human-scale, vascular domains 10 , 11 , 16 . Studies of this nature are better equipped to reveal the subtle, patient-specific, phenomena that exist within complex vasculatures.…”

Section: Introductionmentioning

confidence: 99%

High fidelity blood flow in a patient-specific arteriovenous fistula

McCullough

Coveney

2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

An arteriovenous fistula, created by artificially connecting segments of a patient’s vasculature, is the preferred way to gain access to the bloodstream for kidney dialysis. The increasing power and availability of supercomputing infrastructure means that it is becoming more realistic to use simulations to help identify the best type and location of a fistula for a specific patient. We describe a 3D fistula model that uses the lattice Boltzmann method to simultaneously resolve blood flow in patient-specific arteries and veins. The simulations conducted here, comprising vasculatures of the whole forearm, demonstrate qualified validation against clinical data. Ongoing research to further encompass complex biophysics on realistic time scales will permit the use of human-scale physiological models for basic and clinical medicine.

show abstract

“…On the other hand, further evidence of the efforts being made to increase the efficiency and performance of simulation tools on large supercomputers is provided in the work of McCullough et al [27]. Here the authors highlight several examples as they demonstrate their ongoing developments towards a model for three-dimensional blood flow in coupled, human scale arteries and veins.…”

mentioning

confidence: 99%