Parallel-Prefix Remapping for Efficient Data-Parallel Implementation of Unbalanced Simulations
نویسندگان
چکیده
We present a new algorithm for the eecient data-parallel implementation of spatial individual-or particle-based stochastic (i.e. Monte Carlo) simulations. The potential execution time saving achievable using this algorithm is critically dependent upon the degree of imbalance present in a given application. We have therefore developed a mathematical analysis of the general stochastic diierential equations that govern such models. This analysis provides us with a statistical measure of system behaviour, which can then be used to predict the performance of both previously na ve, and our new implementations.
منابع مشابه
A High Performance Parallel IP Lookup Technique Using Distributed Memory Organization and ISCB-Tree Data Structure
The IP Lookup Process is a key bottleneck in routing due to the increase in routing table size, increasing traıc and migration to IPv6 addresses. The IP address lookup involves computation of the Longest Prefix Matching (LPM), which existing solutions such as BSD Radix Tries, scale poorly when traıc in the router increases or when employed for IPv6 address lookups. In this paper, we describe a ...
متن کاملA High Performance Parallel IP Lookup Technique Using Distributed Memory Organization and ISCB-Tree Data Structure
The IP Lookup Process is a key bottleneck in routing due to the increase in routing table size, increasing traıc and migration to IPv6 addresses. The IP address lookup involves computation of the Longest Prefix Matching (LPM), which existing solutions such as BSD Radix Tries, scale poorly when traıc in the router increases or when employed for IPv6 address lookups. In this paper, we describe a ...
متن کاملEfficient implementation of low time complexity and pipelined bit-parallel polynomial basis multiplier over binary finite fields
This paper presents two efficient implementations of fast and pipelined bit-parallel polynomial basis multipliers over GF (2m) by irreducible pentanomials and trinomials. The architecture of the first multiplier is based on a parallel and independent computation of powers of the polynomial variable. In the second structure only even powers of the polynomial variable are used. The par...
متن کاملImplementation of the direction of arrival estimation algorithms by means of GPU-parallel processing in the Kuda environment (Research Article)
Direction-of-arrival (DOA) estimation of audio signals is critical in different areas, including electronic war, sonar, etc. The beamforming methods like Minimum Variance Distortionless Response (MVDR), Delay-and-Sum (DAS), and subspace-based Multiple Signal Classification (MUSIC) are the most known DOA estimation techniques. The mentioned methods have high computational complexity. Hence using...
متن کاملAdaptive decomposition and remapping algorithms for object-space-parallel direct volume rendering of unstructured grids
Object space (OS) parallelization of an efficient direct volume rendering algorithm for unstructured grids on distributed-memory architectures is investigated. The adaptive OS decomposition problem is modeled as a graph partitioning (GP) problem using an efficient and highly accurate estimation scheme for view-dependent node and edge weighting. In the proposed model, minimizing the cutsize corr...
متن کامل