We desire to permute N items w 0 ... , w N - 1 , in an ultracomputer containing P processing elements (PEs), PE 0 ,... PE P - 1 . Under the assumpution that N+P and that w i e PE i , Schwartz gives the following worst case analyses: The static permutation algorithm requires 4 log P - 3 data communication steps. It is easily seen that for both algorithms the average case behavior closely approximates the worst case. Here we present a data motion algorithm oriented toward average case rather than worst case performance, and supply an argument suggesting that the following average number of data communication steps required is approximately 3 log P. 1. Introduction [UC] introduced the idea of an ultracomputer and reviewed algorithms for two p...
<p>a) Computational performance with increasing number of links and constant number of nodes. The no...
Statistics is a mathematical science pertaining to the collection, analysis, interpretation or expla...
International audienceResearchers and practitioners have for long worked on improving the computatio...
Many papers on parallel random permutation algorithms assume the input size n to be a power of two a...
The traditional permutation multiplication algorithm is now limited by memory latency and not by CPU...
International audienceThe topic of the article is the parametric study of the complexity of algorith...
International audienceWe tackle the feasibility and efficiency of two new parallel algorithms that s...
AbstractTwo “folk theorems” that permeate the parallel computation literature are reconsidered in th...
The permutation generation method is based on starter sets generation under exchange operation and e...
In this paper we give a positive answer to the long-standing problem of finding an in-place sorting...
Technology trends are making the cost of data movement increasingly dominant, both in terms of energ...
Several recent papers have proposed or analyzed optimal algorithms to route all-to-all personalized ...
We study the effect of limited communication throughput on parallel computation in a setting where t...
. Starting from a permutation of f0; : : : ; n \Gamma 1g we compute in parallel with a workload of O...
A mathematical characterization of serially-pruned permutations (SPPs) employed in variable-length p...
<p>a) Computational performance with increasing number of links and constant number of nodes. The no...
Statistics is a mathematical science pertaining to the collection, analysis, interpretation or expla...
International audienceResearchers and practitioners have for long worked on improving the computatio...
Many papers on parallel random permutation algorithms assume the input size n to be a power of two a...
The traditional permutation multiplication algorithm is now limited by memory latency and not by CPU...
International audienceThe topic of the article is the parametric study of the complexity of algorith...
International audienceWe tackle the feasibility and efficiency of two new parallel algorithms that s...
AbstractTwo “folk theorems” that permeate the parallel computation literature are reconsidered in th...
The permutation generation method is based on starter sets generation under exchange operation and e...
In this paper we give a positive answer to the long-standing problem of finding an in-place sorting...
Technology trends are making the cost of data movement increasingly dominant, both in terms of energ...
Several recent papers have proposed or analyzed optimal algorithms to route all-to-all personalized ...
We study the effect of limited communication throughput on parallel computation in a setting where t...
. Starting from a permutation of f0; : : : ; n \Gamma 1g we compute in parallel with a workload of O...
A mathematical characterization of serially-pruned permutations (SPPs) employed in variable-length p...
<p>a) Computational performance with increasing number of links and constant number of nodes. The no...
Statistics is a mathematical science pertaining to the collection, analysis, interpretation or expla...
International audienceResearchers and practitioners have for long worked on improving the computatio...