Toward GPGPU accelerated human electromechanical cardiac simulations.
Bottom Line: Specifically, we port to the GPU a number of components of CHeart--a CPU-based finite element code developed for simulating multi-physics problems.Speedup of up to 72 × compared with SC and 2.6 × compared with MC was also observed for the PDE solve.Using the same human geometry, the GPU implementation of mechanics residual/Jacobian computation provided speedups of up to 44 × compared with SC and 2.0 × compared with MC.
Affiliation: Department of Biomedical Engineering, King's College London, UK.Show MeSH
Related in: MedlinePlus
Mentions: Figure 1 illustrates the hardware interface of CUDA for the Nvidia GPU G80. This parallel single instruction multiple data (SIMD) architecture is endowed with up to 128 cores, where thousands of threads run in parallel. These cores are organized into 16 multiprocessors (SMs), each one having a set of 32-bit registers, constants and texture caches, and 16 KB of on-chip shared memory as fast as local registers (one cycle latency). At any given cycle, each core executes the same instruction on different data (SIMD), and communication between multiprocessors is performed through global memory.
Affiliation: Department of Biomedical Engineering, King's College London, UK.