Skip to main content


Fig. 16 | BMC Systems Biology

Fig. 16

From: Optimisation of an exemplar oculomotor model using multi-objective genetic algorithms executed on a GPU-CPU combination

Fig. 16

CPU/GPU execution times and speedup. a CPU and GPU execution time in seconds (s) vs the number of model integrations. b CPU:GPU execution time ratio (speedup) vs the number of model integrations. The numerical solver used was the implicit mid-point method (time step Δ t=5×10−6 s) and the simulated model time was 6 s. The model parameters were uniformly distributed in the following ranges: α [ 1,1000];β [ 0.001,10];ε [ 0.00001,0.1];γ [ 0.00001,12];α [ 1,1000];β [ 0.001,60]. For each integration, the initial motor error m was set to 2 degrees; all other variables were initially set to 0. In each plot, the dotted line denotes 32000 integrations

Back to article page