Research and Analysis the Performance of NAMD-Molecular Dynamics Simulation Based Multi-core +GPU
Nowadays,the GPU can provide about one order of magnitude accelerated over CPU code and even more,so it can be used in large scale science computing applications.The paper based on NVIDIA Tesla C2050 and dual-core CPU server through use NAMD 2.9 simulates three differences molecule number proteins:Beta2-adrenergic receptor,SET9 and Ubiquitin.Through the comparison and analyses the results of the simulations experiment,we can conclusion that the difference systems of molecule will lead to the difference GPU accelerated.The computing times of four GPU is nearly half of the time used by 1 GPU; and this is especially true in the case of macromolecules.Furthermore,from the GPUs memory utilization rate,the larger the protein system is,the higher the memory use of the GPU is.NVIDIA Tesla C2050 is can satisfy an even larger system simulation.The paper provides the best options for application of this software use the GPU and multi-core framework,a reference to building large scale molecule stimulations platforms,and a solution to science application of large molecule stimulation.
GPU NAMD CUDA molecular stimulation Performance analyses NVIDIA Tesla C2050
Zhang Yang Chen Wen Bo Bai Qi Feng Li lian
Centre of Communication Network,University of Lanzhou,China School of Chemistry and Chemical Engineering,University of Lanzhou,China School of Information Scien
国际会议
杭州
英文
280-283
2013-03-22（万方平台首次上网日期，不代表论文的发表时间）