Variants of Mersenne Twister Suitable for Graphic Processors

Saito, Mutsuo; Matsumoto, Makoto

doi:10.1145/2427023.2427029

Cited by 57 publications

(63 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Para nuestro trabajo hemos usado un generador aleatorio Mersenne Twister [36]. Al comienzo de la ejecución el GPU-DFA define una semilla global con la cual se inicializa una semilla local en cada hilo.…”

Section: Generación De Números Aleatoriosunclassified

Ensamblado de fragmentos de ADN utilizando un novedoso algoritmo de luciérnaga en GPU

Vidal

Olivera

2018

DYNA

View full text Add to dashboard Cite

The Deoxyribonucleic Acid Fragment Assembly Problem (DNA-FAP) consists in reconstruct a DNA chain from a set of fragments taken randomly. Several authors solved the DNA-FAP using different approaches. In general, although it was obtaining good results; the computational time associated is high. The Firefly Algorithm (FA) is a bioinspired model based on the behaviour of fireflies. Considering that FA is a population bioinspired algorithm is possible design a parallel model of itself on Graphics Processing. In this work, a FA especially development for its execution on GPU is presented in order to accelerate the computational process to solve the DNA-FAP. Through several experiments the efficiency of the algorithm and the quality of the results were demonstrated.Keywords: fragment assembly problem; firefly algorithm; graphics processing units; optimization; parallelism. Ensamblado de fragmentos de ADN utilizando un novedoso algoritmo de luciérnaga en GPU ResumenEl problema de ensamblado de fragmentos de cadenas de ácido desoxirribonucleico (Deoxyribonucleic Acid Fragment Assembly Problem, DNA-FAP) consiste en la reconstrucción de cadenas de ADN desde un conjunto de fragmentos tomados aleatoriamente. El DNA-FAP ha sido resuelto por diferentes autores utilizando distintos enfoques. Aunque se obtienen buenos resultados, el tiempo computacional asociado es alto. El algoritmo de luciérnaga (Firefly Algorithm, FA) es un modelo bioinspirado basado en el comportamiento de las luciérnagas. Al ser un algoritmo bioinspirado poblacional es posible generar un modelo paralelo del mismo sobre Unidades de Procesamiento Gráfico (Graphics Processing Units, GPU). En este trabajo un algoritmo de luciérnaga es diseñado especialmente para ser ejecutado sobre una arquitectura GPU de manera tal de acelerar el proceso computacional buscando resolver el DNA-FAP. A través de diferentes experimentos se demuestra la eficiencia computacional y la calidad de los resultados obtenidos.Palabras clave: ensamblado de fragmentos de ADN; algoritmo de luciérnaga; unidades de procesamiento gráfico; optimización; paralelismo. IntroducciónLa necesidad de conocer y predecir mutaciones somáticas y realizar distintos estudios relacionados con el desarrollo de los seres vivos es un tópico de relevancia por su impacto en la medicina para detección y tratamiento de patologías, la investigación forense y el desarrollo de medicamentos. Entre los problemas relacionados con el genoma podemos encontrar el estudio del ensamblado de cadenas de ADN (Deoxyribonucleic Acid Fragment Assembly Problem, DNA-FAP), donde dado un conjunto de cientos o miles de How to cite: Vidal, P.J. and Olivera, A.C., Ensamblado de fragmentos de ADN utilizando un novedoso algoritmo de luciérnaga en GPU. DYNA, 85(204), pp. 108-116, March, 2018. fragmentos de ADN, que pueden contener errores, debemos encontrar la secuencia de ADN original a partir de las permutaciones de los fragmentos que mejor representen a dicha secuencia [17].Existen herramientas que automatizan el secuenciamien...

show abstract

Section: Generación De Números Aleatoriosunclassified

Ensamblado de fragmentos de ADN utilizando un novedoso algoritmo de luciérnaga en GPU

Vidal

Olivera

2018

DYNA

View full text Add to dashboard Cite

show abstract

“…In this study, we use two different generation states to have completely different two 1000-element arrays. One of them generated by MTGP32 pseudorandom sequence generator which is an NVIDIA's adaption of an algorithm proposed by Saito et al [13]. The other state we used is CURAND's default state which generates an array of pseudorandom numbers greater than 2 190 .…”

Section: Exon Shuffling Crossovermentioning

confidence: 99%

A Novel Grouping Genetic Algorithm for the One-Dimensional Bin Packing Problem on GPU

Ozcan

Dökeroğlu

Çoşar

et al. 2016

Communications in Computer and Information Science

View full text Add to dashboard Cite

Abstract. One-dimensional Bin Packing Problem (1D-BPP) is a challenging NP-Hard combinatorial problem which is used to pack finite number of items into minimum number of bins. Large problem instances of the 1D-BPP cannot be solved exactly due to the intractable nature of the problem. In this study, we propose an efficient Grouping Genetic Algorithm (GGA) by harnessing the power of the Graphics Processing Unit (GPU) using CUDA. The time consuming crossover and mutation processes of the GGA are executed on the GPU by increasing the evaluation times significantly. The obtained experimental results on 1,238 benchmark 1D-BPP instances show that our proposed algorithm has a high performance and is a scalable algorithm with its high speed fitness evaluation ability. Our proposed algorithm can be considered as one of the best performing algorithms with its 66 times faster computation speed that enables to explore the search space more effectively than any of its counterparts.

show abstract

“…Another approach to produce multiple streams uses a different RNG or a different set of RNG parameters for each stream (Mascagni andSrinivasan 2000, Saito andMatsumoto 2013). This is generally much less convenient than using the same RNG (and same code) for all streams, for many reasons (L 'Ecuyer et al 2015).…”

Section: Making and Managing Multiple Streamsmentioning

confidence: 99%

Random number generation with multiple streams for sequential and parallel computing

L’Ecuyer¹

2015

2015 Winter Simulation Conference (WSC)

View full text Add to dashboard Cite

We provide a review of the state of the art on the design and implementation of random number generators (RNGs) for simulation, on both sequential and parallel computing environments. We focus on the need for multiple streams and substreams of random numbers, explain how they can be constructed and managed, review software libraries that offer them, and illustrate their usefulness via examples. We also review the basic quality criteria for good random number generators and their theoretical and empirical testing.

show abstract

Variants of Mersenne Twister Suitable for Graphic Processors

Cited by 57 publications

References 21 publications

Ensamblado de fragmentos de ADN utilizando un novedoso algoritmo de luciérnaga en GPU

Ensamblado de fragmentos de ADN utilizando un novedoso algoritmo de luciérnaga en GPU

A Novel Grouping Genetic Algorithm for the One-Dimensional Bin Packing Problem on GPU

Random number generation with multiple streams for sequential and parallel computing

Contact Info

Product

Resources

About