Publicaciones

@article{barrientos_gpu-based_2017,
title = {GPU-based exhaustive algorithms processing kNN queries},
author = {Ricardo J Barrientos and Fabricio Millaguir and José L Sánchez and Enrique Arias},
url = {http://link.springer.com/10.1007/s11227-017-2110-y},
doi = {10.1007/s11227-017-2110-y},
issn = {0920-8542},
volume = {73},
number = {10},
pages = {4611--4634},
abstract = {Efficient kNN search, or k-nearest neighbors search, is useful, among other fields, in multimedia information retrieval, data mining and pattern recognition problems. A distance function determines how similar the objects are to a given kNN query object. As finding the distance between any given pair of objects (i.e., high-dimensional vectors) is known to be a computationally expensive operation, using parallel computation techniques is an effective way of reducing running times to acceptable values in large databases. In the present work, we offer novel GPU approaches to solving kNN (k-nearest neighbor) queries using exhaustive algorithms based on the Selection Sort, Quicksort and state-of-the-art algorithms. We show that the best approach depends on the k value of the kNN query and achieve a speedup up to 86.4× better than the sequential counterpart. We also propose a multi-core algorithm to be used as reference for the experiments and a hybrid algorithm which combines the proposed algorithms with a state-of-the-art heaps-based method, in which the best performance is obtained with high k values. We also extend our algorithms to be able to deal with large databases that do not fit in GPU memory and whose performance does not deteriorate as database size increases.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}

@article{peralta_vibrational_2017,
title = {Vibrational properties of Cu3XY4 sulvanites (X = Nb, Ta, and V; and Y = S, and Se) by ab initio molecular dynamics},
author = {Joaquín Peralta and Camilo Valencia Balvín},
url = {http://link.springer.com/10.1140/epjb/e2017-80050-7},
doi = {10.1140/epjb/e2017-80050-7},
issn = {1434-6028},
volume = {90},
number = {9},
pages = {177},
abstract = {In this work, we present a structural and dynamic characterisation of six different types of sulvanites Cu 3 XY 4 with X = Nb, V and Ta, and Y = S and Se. These materials have been the subject of intense study in recent times primarily as potential candidates for solar cell devices, as well as for their enhanced opto-electrical properties. Here, by means of first-principles calculations, we study the structural and dynamic behaviour of these materials at different temperatures, which is important for use of these materials in high-temperature conditions. In this work the dynamic and structural properties are studied using the Density Functional Theory technique. The simulations were performed at four different temperatures, ranging from room temperature to ∼ 1500 K. By using first-principles molecular dynamics in the microcanonical ensemble, we are able to determine the vibrational spectra of these sulvanites. With this information we report for the first time the partial vibrational density of states of these structures at different temperatures. With these results we determine the vibrational properties of the basic building blocks of those sulvanites and their dynamic behaviour under temperature effects. We also show that the building blocks that which make up these structures, remain stable as the temperature increases.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}

@article{rabanal-le_exploring_nodate,
title = {Exploring the potential energy surface of small lead clusters using the gradient embedded genetic algorithm and an adequate treatment of relativistic effects},
author = {Walter A Rabanal-Lé and William Tiznado and Edison Osorio and Franklin Ferraro},
doi = {10.1039/c7ra11449d},
abstract = {It is a well-known fact that theoretical methodologies play a crucial role to assure an adequate structural assignment of gas-phase clusters. Particularly, in heavy-element containing clusters the inclusion of relativistic effects (scalar and spin–orbit coupling) can significantly affect their chemistry. Therefore, these effects become the keystone on their structural determination. In our work, the way in which relativistic effects were treated, as well as their influence in the process of an adequate identification of lowest-energy isomer (the global minima – " GM " – energy structure), were evaluated in small lead clusters. The potential energy surfaces of small Pb n (n ¼ 3–10) clusters was explored by means of the gradient embedded genetic algorithm program (GEGA). Subsequently, the most stable isomers were re-optimized incorporating relativistic effects through two different approximations: (i) using relativistic effective core potentials (RECPs) or pseudopotentials, which mimics the scalar and spin–orbit coupling relativistic effects (SR and SO) of the core electrons; and (ii) using relativistic Hamiltonians (with proper all-electron basis sets), like, the zeroth-order regular approximation (ZORA) to the Dirac equation, in which the scalar (SR) and spin–orbit coupling (SOC) relativistic effects were also included. The results evidence that methodologies including SOC effect allow to identify the GM energy structure correctly in all the studied cases. Besides, the GEGA algorithm, using a modest RECP, provides good initial structures that become GM after re-optimization at the SOC level.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}