Marcos Novalbos - Academia.edu (original) (raw)
Uploads
Papers by Marcos Novalbos
Celebrado los dias 21-23 de Marzo del 2012, Universidad Rey Juan Carlos, Mostoles, Madrid
Parallel architectures, in the form of multi-core or multiple computers, have produced a major im... more Parallel architectures, in the form of multi-core or multiple computers, have produced a major impact in the field of information technology. GPU devices, as an extreme example of parallel architectures, have been adapted to enable generic computation in massively parallel architectures. Molecular dynamics is a problem that fits perfectly such architectures, as it relies on the computation of many similar interactions between atoms. Moreover, large molecular systems require resources that exceed those available in a single computer, even multi-GPU computers. Therefore, the ideal architecture to simulate molecular dynamics is a distributed multi-GPU cluster, which consists of multiple interconnected computers with one or more GPUs each. A molecular dynamics simulation usually needs days, and even weeks of computation time to produce results that represent only a few microseconds of atom interactions. In contrast, distributed multi-GPU clusters allows us to develop an efficient and sc...
Journal of Parallel and Distributed Computing, 2021
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Lecture Notes in Computer Science, 2014
Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems by ... more Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems by modeling the pairwise interaction forces between all atoms. Molecular systems are subject to slowly decaying electrostatic potentials, which turn molecular dynamics into an n-body problem. In this paper, we present a parallel and scalable solution to compute long-range molecular forces, based on the multilevel summation method (MSM). We first demonstrate an optimization of MSM that replaces 3D convolutions with FFTs, and we achieve a single-GPU performance comparable to the particle mesh Ewald (PME) method, the de facto standard for long-range molecular force computation. But most importantly, we propose a distributed MSM that avoids the scalability difficulties of PME. Our distributed solution is based on a spatial partitioning of the MSM multilevel grid, together with massively parallel algorithms for interface update and synchronization. We demonstrate the scalability of our approach on an on-board multi-GPU platform.
Lecture Notes in Computer Science, 2013
Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems. Th... more Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems. These simulations suffer a large computational complexity that leads to simulation times of several weeks in order to recreate just a few microseconds of a molecule's motion even on high-performance computing platforms. In recent years, state-ofthe-art molecular dynamics algorithms have benefited from the parallel computing capabilities of multicore systems, as well as GPUs used as co-processors. In this paper we present a parallel molecular dynamics algorithm for on-board multi-GPU architectures. We parallelize a stateof-the-art molecular dynamics algorithm at two levels. We employ a spatial partitioning approach to simulate the dynamics of one portion of a molecular system on each GPU, and we take advantage of direct communication between GPUs to transfer data among portions. We also parallelize the simulation algorithm to exploit the multi-processor computing model of GPUs. Most importantly, we present novel parallel algorithms to update the spatial partitioning and set up transfer data packages on each GPU. We demonstrate the feasibility and scalability of our proposal through a comparative study with NAMD, a well known parallel molecular dynamics implementation.
Higher resolution meshes should be used in graphics applications to make them as realistic as the... more Higher resolution meshes should be used in graphics applications to make them as realistic as they can. However, they imply a high computational. Several approaches have been built to solve collision detection, although most of them do not take into account this feature. This paper presents a scalable parallel algorithm for collision detection designed for working with high resolution meshes. The algorithm is based on distributed memory architectures taking advantage of their benefits and overcoming their drawbacks.
Celebrado los dias 21-23 de Marzo del 2012, Universidad Rey Juan Carlos, Mostoles, Madrid
Parallel architectures, in the form of multi-core or multiple computers, have produced a major im... more Parallel architectures, in the form of multi-core or multiple computers, have produced a major impact in the field of information technology. GPU devices, as an extreme example of parallel architectures, have been adapted to enable generic computation in massively parallel architectures. Molecular dynamics is a problem that fits perfectly such architectures, as it relies on the computation of many similar interactions between atoms. Moreover, large molecular systems require resources that exceed those available in a single computer, even multi-GPU computers. Therefore, the ideal architecture to simulate molecular dynamics is a distributed multi-GPU cluster, which consists of multiple interconnected computers with one or more GPUs each. A molecular dynamics simulation usually needs days, and even weeks of computation time to produce results that represent only a few microseconds of atom interactions. In contrast, distributed multi-GPU clusters allows us to develop an efficient and sc...
Journal of Parallel and Distributed Computing, 2021
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Lecture Notes in Computer Science, 2014
Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems by ... more Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems by modeling the pairwise interaction forces between all atoms. Molecular systems are subject to slowly decaying electrostatic potentials, which turn molecular dynamics into an n-body problem. In this paper, we present a parallel and scalable solution to compute long-range molecular forces, based on the multilevel summation method (MSM). We first demonstrate an optimization of MSM that replaces 3D convolutions with FFTs, and we achieve a single-GPU performance comparable to the particle mesh Ewald (PME) method, the de facto standard for long-range molecular force computation. But most importantly, we propose a distributed MSM that avoids the scalability difficulties of PME. Our distributed solution is based on a spatial partitioning of the MSM multilevel grid, together with massively parallel algorithms for interface update and synchronization. We demonstrate the scalability of our approach on an on-board multi-GPU platform.
Lecture Notes in Computer Science, 2013
Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems. Th... more Molecular dynamics simulations allow us to study the behavior of complex biomolecular systems. These simulations suffer a large computational complexity that leads to simulation times of several weeks in order to recreate just a few microseconds of a molecule's motion even on high-performance computing platforms. In recent years, state-ofthe-art molecular dynamics algorithms have benefited from the parallel computing capabilities of multicore systems, as well as GPUs used as co-processors. In this paper we present a parallel molecular dynamics algorithm for on-board multi-GPU architectures. We parallelize a stateof-the-art molecular dynamics algorithm at two levels. We employ a spatial partitioning approach to simulate the dynamics of one portion of a molecular system on each GPU, and we take advantage of direct communication between GPUs to transfer data among portions. We also parallelize the simulation algorithm to exploit the multi-processor computing model of GPUs. Most importantly, we present novel parallel algorithms to update the spatial partitioning and set up transfer data packages on each GPU. We demonstrate the feasibility and scalability of our proposal through a comparative study with NAMD, a well known parallel molecular dynamics implementation.
Higher resolution meshes should be used in graphics applications to make them as realistic as the... more Higher resolution meshes should be used in graphics applications to make them as realistic as they can. However, they imply a high computational. Several approaches have been built to solve collision detection, although most of them do not take into account this feature. This paper presents a scalable parallel algorithm for collision detection designed for working with high resolution meshes. The algorithm is based on distributed memory architectures taking advantage of their benefits and overcoming their drawbacks.