scholarly journals PAVER

2021 ◽  
Vol 18 (3) ◽  
pp. 1-26
Author(s):  
Devashree Tripathy ◽  
Amirali Abdolrashidi ◽  
Laxmi Narayan Bhuyan ◽  
Liang Zhou ◽  
Daniel Wong

The massive parallelism present in GPUs comes at the cost of reduced L1 and L2 cache sizes per thread, leading to serious cache contention problems such as thrashing. Hence, the data access locality of an application should be considered during thread scheduling to improve execution time and energy consumption. Recent works have tried to use the locality behavior of regular and structured applications in thread scheduling, but the difficult case of irregular and unstructured parallel applications remains to be explored. We present PAVER , a P riority- A ware V ertex schedul ER , which takes a graph-theoretic approach toward thread scheduling. We analyze the cache locality behavior among thread blocks ( TBs ) through a just-in-time compilation, and represent the problem using a graph representing the TBs and the locality among them. This graph is then partitioned to TB groups that display maximum data sharing, which are then assigned to the same streaming multiprocessor by the locality-aware TB scheduler. Through exhaustive simulation in Fermi, Pascal, and Volta architectures using a number of scheduling techniques, we show that PAVER reduces L2 accesses by 43.3%, 48.5%, and 40.21% and increases the average performance benefit by 29%, 49.1%, and 41.2% for the benchmarks with high inter-TB locality.


10.29007/l442 ◽  
2018 ◽  
Author(s):  
Jure Zevnik ◽  
Daniel Kozelj

We present a method for partitioning Water Distribution Networks (WDNs) into District Metered Areas (DMAs) by using a spectral graph partitioning algorithm. The effectiveness of DMA design was tested for selected edge weights and multiple numbers of established DMAs. The presented method includes a novel graph theoretic approach to determine and evaluate only relevant combinations of DMA connection. It was tested on a real-life case study for which several different solutions were generated and evaluated against their hydraulic performance. The optimal solution, i.e. design of DMAs, was selected regarding the quality of partition and the cost of WDN segmentation, since hydraulic adequacy was met for all cases where partitioning resulted in connected subgraphs.



Author(s):  
C.D. YANG ◽  
D.T. LEE ◽  
C.K. WONG

We consider the problem of finding a rectilinear path between two designated points in the presence of rectilinear obstacles subject to various optimization functions in terms of the number of bends and the total length of the path. Specifically we are interested in finding a minimum bend shortest path, a shortest minimum bend path or a least-cost path where the cost is defined as a function of both the length and the number of bends of the path. We provide a unified approach by constructing a path-preserving graph. guaranteed to preserve all these three kinds of paths and give an O(K+e log e) algorithm to find them, where e is the total number of obstacle edges, and K is the number of intersections between tracks from extreme point and other tracks (defined in the text). K is bounded by O(et), where t is the number of extreme edges. In particular, if the obstacles are rectilinearly convex, then K is O(ne), where n is the number of obstacles. Extensions are made to find a shortest path with a bounded number of bends and a minimum-bend path with a bounded length. When a source point and obstacles are pre-given, queries for the assorted paths from the source to given points can be handled in O( log e+k) time after O(K+e log e) preprocessing, where k is the size of the goal path. The trans-dichotomous algorithm of Fredman and Willard8 and the running time for these problems are also discussed.





2021 ◽  
Vol 10 (1) ◽  
pp. 91-108
Author(s):  
Aziz Ullah Awan ◽  
Attia Sharif ◽  
Kashif Ali Abro ◽  
Muhammad Ozair ◽  
Takasar Hussain

Abstract The square-root dynamics of smoking model with cravings to smoke, in which square root of potential smokers and smokers is the interaction term, has been studied. We categorized net population in four different chambers: non-smokers/potential smokers, smokers/infected people, non-permanent smokers/temporary quitters and the permanent quitters. By dynamical systems approach, we analyzed our model. Moreover, for proving the unique equilibrium point to be globally stable, we took help of graph theoretic approach. The sensitivity analysis of the model is performed through the diseased classes effectively to design reliable, robust and stable control strategies. The model is designed like optimal control trouble to find out importance of various control actions on our system that are insisted by the sensitivity analysis. We have applied two controls, which are the awareness campaign through the media transmission to control the potential smokers and temporary quit smokers to become smokers and the treatment of smokers. Analytical and numerical methods are utilized for ensuring presence of these two control actions.



Genetics ◽  
2003 ◽  
Vol 165 (4) ◽  
pp. 2235-2247
Author(s):  
Immanuel V Yap ◽  
David Schneider ◽  
Jon Kleinberg ◽  
David Matthews ◽  
Samuel Cartinhour ◽  
...  

AbstractFor many species, multiple maps are available, often constructed independently by different research groups using different sets of markers and different source material. Integration of these maps provides a higher density of markers and greater genome coverage than is possible using a single study. In this article, we describe a novel approach to comparing and integrating maps by using abstract graphs. A map is modeled as a directed graph in which nodes represent mapped markers and edges define the order of adjacent markers. Independently constructed graphs representing corresponding maps from different studies are merged on the basis of their common loci. Absence of a path between two nodes indicates that their order is undetermined. A cycle indicates inconsistency among the mapping studies with regard to the order of the loci involved. The integrated graph thus produced represents a complete picture of all of the mapping studies that comprise it, including all of the ambiguities and inconsistencies among them. The objective of this representation is to guide additional research aimed at interpreting these ambiguities and inconsistencies in locus order rather than presenting a “consensus order” that ignores these problems.



2020 ◽  
Vol 1706 ◽  
pp. 012115
Author(s):  
P Sangeetha ◽  
M Shanmugapriya ◽  
R Sundareswaran ◽  
K Sowmya ◽  
S Srinidhi




Sign in / Sign up

Export Citation Format

Share Document