Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 4
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Real-space analysis of branch point motion in architecturally complex polymers

  • Bačová, Petra
  • Moreno Segurado, Ángel J.
By means of large-scale molecular dynamics simulations, we investigate branch point motion in pure branched polymers and in mixtures of stars and linear chains. We perform a purely geometrical density-based cluster analysis of the branch point trajectories and identify regions of strong localization (traps). Our results demonstrate that the branch point motion can be described as the motion over a network of traps at the time scales corresponding to the reptation regime. Residence times within the traps are broadly distributed, even extending to times much longer than the side-arm relaxation time. The distributions of distances between consecutively visited traps are very similar for all the investigated branched polymers, even though tube dilation is much stronger in the star/linear mixtures than in the pure branched systems. Our analysis suggests that the diffusivity of the branch point introduced by hierarchical models must be understood as a parameter to account for the effective friction associated with the relaxed side arm, more than the description of a hopping process with a precise time scale., We acknowledge support from projects FP7-PEOPLE-2007-1-1-ITN (DYNACOP, EU), MAT2012-31088 (Spain), and IT654-13 (GV, Spain). We acknowledge the programs PRACE, HPC-Europa2 and ESMI (EU), and ICTS (Spain) for generous allocation of CPU time at GENCI (France), HLRS and FZJ-JSC (Germany), and CESGA (Spain)., Peer Reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Tesis doctoral (DoctoralThesis)

Computational investigation of chain dynamics in architecturally complex polymers

  • Bačová, Petra
We investigate the chain dynamics in the polymer melts with complex architecture by means of molecular dynamics simulations. Our study is focused on the following architecturally complex polymers: T- and Y-shaped asymmetric stars, symmetric stars, mixtures of Tshaped asymmetric stars and linear chains, H-polymers, combs and Cayley trees. Dynamics in these architectures is strongly influenced by the presence of one or more branchpoints. The overall chain dynamics in branched structures is slowed down comparing to the linear chain and the relaxation of these materials extend over several time decades. Extensive molecular dynamics simulations allow us to study the relaxation processes ocurring in the branched polymer melts at the molecular level. We pay particular attention to the role of the branchpoint in the dynamics of these systems. Our simulations reveal details about the branchpoint motion that can be further compared to the theoretical hypotheses, experimental data and finally introduced in the predictions of the viscoelastic properties of the industrially produced materials. The time evolutions of the mean squared displacements of the particular molecular segments confirm that the arm retraction is the main relaxation mechanism in symmetric systems, i.e. symmetric stars and Cayley trees. In these systems the branchpoint remains localized during the whole simulation time. We study the role of constraint release on the branchpoint dynamics and compare the simulation results with a theoretical model. The fluctuations of the branchpoint at time scales smaller than the Rouse time τR are affected by the early tube dilation process that leads to a weaker branchpoint localization than expected. After the incorporation of the early and late tube dilation processes quantified from the simulation data into the theoretical model, we were able to fully describe the branchpoint dynamics at the times smaller than τR. After the relaxation of the short side arms in the asymmetric structures the arms act as sources of friction and the whole molecule can be described as an effective linear chain. We studied the diffusion of the branchpoint after the arm relaxation. The calculation of the diffusion constant involves the knowledge about the arm relaxation, dilution of the tube and the friction related to the reptation of the molecule. We estimated these observables from the simulations and tested theoretical hypotheses used for the prediction of the branchpoint diffusive behaviour in experimental studies of branched polymer melts. We perform a detailed analysis of the branchpoint trajectories and present a robust method for finding regions of strong localization. We characterize the time and length scales for the branchpoint motion between traps of localization, and discuss the consequences for the interpretation of the long-time branchpoint motion dynamics proposed by hierarchical tube models., This thesis would have not been possible without the financial support from projects FP7-PEOPLE-2007-1-1-ITN (DYNACOP, EU), MAT2012-31088 (Spain) and IT654-13 (GV, Spain). I would like to acknowledge the programs PRACE, HPC-Europa2 and ESMI (EU), and ICTS (Spain) for generous allocation of CPU time at GENCI (France), HLRS and FZJ-JSC (Germany) and CESGA (Spain)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Algorithm 942: Semi-Stencil

  • de la Cruz, Raúl
  • Araya-Polo, Mauricio
Finite Difference (FD) is a widely used method to solve Partial Differential Equations (PDE). PDEs are the core of many simulations in different scientific fields, such as geophysics, astrophysics, etc. The typical FD solver performs stencil computations for the entire computational domain, thus solving the differential operators. In general terms, the stencil computation consists of a weighted accumulation of the contribution of neighbor points along the cartesian axis. Therefore, optimizing stencil computations is crucial in reducing the application execution time. Stencil computation performance is bounded by two main factors: the memory access pattern and the inefficient reuse of the accessed data. We propose a novel algorithm, named Semi-stencil, that tackles these two problems. The main idea behind this algorithm is to change the way in which the stencil computation progresses within the computational domain. Instead of accessing all required neighbors and adding all their contributions at once, the Semi-stencil algorithm divides the computation into several updates. Then, each update gathers half of the axis neighbors, partially computing at the same time the stencil in a set of closely located points. As Semi-stencil progresses through the domain, the stencil computations are completed on precomputed points. This computation strategy improves the memory access pattern and efficiently reuses the accessed data. Our initial target architecture was the Cell/B.E., where the Semi-stencil in a SPE was 44% faster than the naive stencil implementation. Since then, we have continued our research on emerging multicore architectures in order to assess and extend this work on homogeneous architectures. The experiments presented combine the Semi-stencil strategy with space- and time-blocking algorithms used in hierarchical memory architectures. Two x86 (Intel Nehalem and AMD Opteron) and two POWER (IBM POWER6 and IBM BG/P) platforms are used as testbeds, where the best improvements for a 25-point stencil range from 1.27 to 1.76× faster. The results show that this novel strategy is a feasible optimization method which may be integrated into auto-tuning frameworks. Also, since all current architectures are multicore based, we have introduced a brief section where scalability results on IBM POWER7-, Intel Xeon-, and MIC-based systems are presented. In a nutshell, the algorithm scales as well as or better than other stencil techniques. For instance, the scalability of Semi-stencil on MIC for a certain testcase reached 93.8 × over 244 threads., The authors would like to express their gratitude to the Partnership for Advanced Computing in Europe (PRACE) for their grant of computing time on the Huygens, Juropa, Jugene and Louhi supercomputing facilities. The PRACE project receives funding from the EU’s Seventh Framework Programme (FP7/2007-2013) under grant agreement no. RI-211528. The IBM PS701 BladeCenter used in this paper was funded by the Spanish Ministry of Science and Innovation under grant 2009501052 awarded to Francisco J. Cazorla. The authors also thank to Roberto, Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Algorithm 942: Semi-Stencil

  • de la Cruz, Raúl
  • Araya-Polo, Mauricio
Finite Difference (FD) is a widely used method to solve Partial Differential Equations (PDE). PDEs are the core of many simulations in different scientific fields, such as geophysics, astrophysics, etc. The typical FD solver performs stencil computations for the entire computational domain, thus solving the differential operators. In general terms, the stencil computation consists of a weighted accumulation of the contribution of neighbor points along the cartesian axis. Therefore, optimizing stencil computations is crucial in reducing the application execution time. Stencil computation performance is bounded by two main factors: the memory access pattern and the inefficient reuse of the accessed data. We propose a novel algorithm, named Semi-stencil, that tackles these two problems. The main idea behind this algorithm is to change the way in which the stencil computation progresses within the computational domain. Instead of accessing all required neighbors and adding all their contributions at once, the Semi-stencil algorithm divides the computation into several updates. Then, each update gathers half of the axis neighbors, partially computing at the same time the stencil in a set of closely located points. As Semi-stencil progresses through the domain, the stencil computations are completed on precomputed points. This computation strategy improves the memory access pattern and efficiently reuses the accessed data. Our initial target architecture was the Cell/B.E., where the Semi-stencil in a SPE was 44% faster than the naive stencil implementation. Since then, we have continued our research on emerging multicore architectures in order to assess and extend this work on homogeneous architectures. The experiments presented combine the Semi-stencil strategy with space- and time-blocking algorithms used in hierarchical memory architectures. Two x86 (Intel Nehalem and AMD Opteron) and two POWER (IBM POWER6 and IBM BG/P) platforms are used as testbeds, where the best improvements for a 25-point stencil range from 1.27 to 1.76× faster. The results show that this novel strategy is a feasible optimization method which may be integrated into auto-tuning frameworks. Also, since all current architectures are multicore based, we have introduced a brief section where scalability results on IBM POWER7-, Intel Xeon-, and MIC-based systems are presented. In a nutshell, the algorithm scales as well as or better than other stencil techniques. For instance, the scalability of Semi-stencil on MIC for a certain testcase reached 93.8 × over 244 threads., The authors would like to express their gratitude to the Partnership for Advanced Computing in Europe (PRACE) for their grant of computing time on the Huygens, Juropa, Jugene and Louhi supercomputing facilities. The PRACE project receives funding from the EU’s Seventh Framework Programme (FP7/2007-2013) under grant agreement no. RI-211528. The IBM PS701 BladeCenter used in this paper was funded by the Spanish Ministry of Science and Innovation under grant 2009501052 awarded to Francisco J. Cazorla. The authors also thank to Roberto, Peer Reviewed, Postprint (author's final draft)