SUBCUBE ALLOCATION IN HYPERCUBE COMPUTERS

被引:30
作者
DUTT, S
HAYES, JP
机构
[1] Advanced Computer Architecture Laboratory, Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor
关键词
ALLOCATION ALGORITHMS; COALESCING ALGORITHMS; HYPERCUBE COMPUTERS; HYPERCUBE FRAGMENTATION; MULTIPROCESSORS; NP-COMPLETE PROBLEMS; PROCESSOR ALLOCATION; SUBCUBE PACKING;
D O I
10.1109/12.76413
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
In hypercube computers that support a multiuser environment, it is important for the operating system to be able to allocate subcubes of different dimensions. Previously proposed subcube allocation schemes, such as the buddy strategy, can fragment the hypercube excessively. We present a precise characterization of the subcube allocation problem and develop a general methodology to solve it. New subcube allocation and coalescing algorithms are described that have the goal of minimizing fragmentation. The concept of a maximal set of subcubes (MSS), which is useful in making allocations that result in a tightly packed hypercube, is introduced. The problems of allocating subcubes and of forming an MSS are formulated as decision problems, and shown to be NP-hard. We prove analytically that the buddy strategy is optimal under restricted conditions, and then show using simulation that its performance is actually poor under more realistic conditions. We suggest a heuristic procedure for efficiently coalescing a released cube with the existing free cubes. This coalescing approach is coupled with a simple best-fit allocation scheme to form the basis of a class of MSS-based strategies that give a substantial performance (hit ratio) improvement over the buddy strategy. Finally, we present simulation results comparing several different allocation and coalescing strategies, which show that our MSS-based schemes provide a marked performance improvement over previous techniques.
引用
收藏
页码:341 / 352
页数:12
相关论文
共 50 条
  • [41] Multi-Constraint Multi-Processor Resource Allocation
    Behrouzian, Amir R. B.
    Goswami, Dip
    Basten, Twan
    Geilen, Marc
    Ara, Hadi Alizadeh
    PROCEEDINGS INTERNATIONAL CONFERENCE ON EMBEDDED COMPUTER SYSTEMS - ARCHITECTURES, MODELING AND SIMULATION (SAMOS XV), 2015, : 338 - 346
  • [42] An effective processor allocation strategy for multiprogrammed shared-memory multiprocessors
    Yue, KK
    Lilja, DJ
    IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, 1997, 8 (12) : 1246 - 1258
  • [43] Performance evaluation of task migration in contiguous allocation for mesh interconnection topology
    Mahnaz Rafie
    Ahmad Khademzadeh
    Akram Reza
    Midia Reshadi
    The Journal of Supercomputing, 2016, 72 : 1660 - 1677
  • [44] Optimization of data distribution and processor allocation problem using simulated annealing
    Onbasioglu, E
    Özdamar, L
    JOURNAL OF SUPERCOMPUTING, 2003, 25 (03) : 237 - 253
  • [45] Processor allocation and task scheduling of matrix chain products on parallel systems
    Lee, H. (heejo@ahnlab.com), 1600, Institute of Electrical and Electronics Engineers Computer Society (14): : 394 - 407
  • [46] Comparing processor allocation strategies in multiprogrammed shared-memory multiprocessors
    Yue, KK
    Lilja, DJ
    JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 1998, 49 (02) : 245 - 258
  • [47] Some new results in the complexity of allocation and binding in Data Path Synthesis
    Mandal, CA
    Chakrabarti, PP
    Ghose, S
    COMPUTERS & MATHEMATICS WITH APPLICATIONS, 1998, 35 (10) : 93 - 105
  • [48] Comparison of Allocation Algorithms in Mesh Oriented Structures for Different Scheduling Techniques
    Bodzon, Bartosz
    Koszalka, Leszek
    Pozniak-Koszalka, Iwona
    Kasprzak, Andrzej
    COMPUTATIONAL COLLECTIVE INTELLIGENCE - TECHNOLOGIES AND APPLICATIONS, PT II, 2012, 7654 : 223 - 232
  • [49] Performance evaluation of task migration in contiguous allocation for mesh interconnection topology
    Rafie, Mahnaz
    Khademzadeh, Ahmad
    Reza, Akram
    Reshadi, Midia
    JOURNAL OF SUPERCOMPUTING, 2016, 72 (04) : 1660 - 1677
  • [50] Hyper-heuristic method for processor allocation in parallel tasks scheduling
    Yildiz, Gulcin
    Sevilgen, Fatih Erdogan
    CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2023, 35 (24)