Skip to main content
Figure 12 | Theoretical Biology and Medical Modelling

Figure 12

From: Developing a multiscale, multi-resolution agent-based brain tumor model by graphics processing units

Figure 12

(a) Computation time of GPU -based multiscale model and sequential multiscale model. The x-axis represents the high-resolution lattice size and the y-axis represents the computation time (logarithmic scale with base 10) in milliseconds. The blue bar represents the computation time of the sequential multiscale model and the red bar represents the computation time of GPU-based multiscale model. The number on each red bar indicates the speed of the parallelized algorithm divided by the speed of the sequential algorithm. (b) Computation time of GPU -based multi-resolution and multiscale model and GPU -based multiscale model. The x axis represents the high-resolution lattice size and y axis represents the computation time (logarithmic scale with base 10) in milliseconds. The blue bar represents the computation time of GPU-based multiscale model and the red bar represents the computation time of the GPU-based multiscale and multi-resolution model. The number on the red bar indicates the speed-up of the GPU-based multiscale model. (c) Computation time of GPU -based multi-resolution and multiscale model with parallel ODE and PDE modules and with only parallel PDE module. The x axis represents the total cell number and the y axis represents the computation time in milliseconds. The blue bar represents the computation time of the GPU-based multi-resolution and multi-scale model with only the PDE solver parallelized. The red bar represents the computation time of the GPU-based multi-resolution and multiscale model with both the ODE and PDE modules parallelized. The number on the red bar indicates the speed-up of the GPU-based multiscale and multi-resolution model with only the PDE module parallelized.

Back to article page