POSTER: Scheduling HPC Workloads on Heterogeneous-ISA Architectures

被引:3
|
作者
Karaoui, Mohamed L. [1 ]
Carno, Anthony [1 ]
Lyerly, Rob [1 ]
Kim, Sang-Hoon [1 ]
Olivier, Pierre [1 ]
Min, Changwoo [1 ]
Ravindran, Binoy [1 ]
机构
[1] Virginia Tech, Blacksburg, VA 24061 USA
关键词
D O I
10.1145/3293883.3295717
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
In this paper, we investigate the effectiveness of multiprocessor architectures with ISA-different cores for executing HPC workloads. Our envisioned design point in the heterogeneous architecture space is one with multiple cache-coherency domains, with each domain hosting cores of a different ISA and no coherency between domains. We prototype such an architecture using an Intel Xeon x86-64 server and a Cavium ThunderX ARMv8 server, interconnected using a high-speed network fabric. We design, implement, and evaluate policies for scheduling HPC applications with the goal of maximizing workload makespan. Our results reveal that such an architecture is most effective for workloads that exhibit diverse execution times on ISA-different CPUs, with gains exceeding 60% over ISA-homogeneous architectures. Furthermore, cross-ISA execution migration can yield gains up to 38%.
引用
收藏
页码:409 / 410
页数:2
相关论文
共 50 条
  • [21] Scheduling divisible workloads on heterogeneous platforms
    Beaumont, O
    Legrand, A
    Robert, Y
    PARALLEL COMPUTING, 2003, 29 (09) : 1121 - 1152
  • [22] Performance Modelling of Heterogeneous ISA Multicore Architectures
    Boran, Nirmal Kumar
    Meghwal, Rameshwar Prasad
    Sharma, Kuldeep
    Kumar, Binod
    Singh, Virendra
    PROCEEDINGS OF 2016 IEEE EAST-WEST DESIGN & TEST SYMPOSIUM (EWDTS), 2016,
  • [23] Multiple Objective Scheduling of HPC Workloads Through Dynamic Prioritization
    Simon, Tyler A.
    Phuong Nguyen
    Halem, Milton
    HIGH PERFORMANCE COMPUTING SYMPOSIUM 2013 (HPC 2013) - 2013 SPRING SIMULATION MULTI-CONFERENCE (SPRINGSIM'13), 2013, 45 (06): : 105 - 112
  • [24] Mapping and scheduling for heterogeneous architectures
    Ramos-Hernandez, DN
    Tokhi, MO
    Bass, JM
    MICROPROCESSORS AND MICROSYSTEMS, 1999, 23 (01) : 7 - 23
  • [25] SIMULATION OF HPC JOB SCHEDULING AND LARGE-SCALE PARALLEL WORKLOADS
    Abu Obaida, Mohammad
    Liu, Jason
    2017 WINTER SIMULATION CONFERENCE (WSC), 2017, : 920 - 931
  • [26] Exposing the Locality of Heterogeneous Memory Architectures to HPC Applications
    Goglin, Brice
    MEMSYS 2016: PROCEEDINGS OF THE INTERNATIONAL SYMPOSIUM ON MEMORY SYSTEMS, 2016, : 30 - 39
  • [27] Adaptive statistical scheduling of divisible workloads in heterogeneous systems
    Gonzalez-Velez, Horacio
    Cole, Murray
    JOURNAL OF SCHEDULING, 2010, 13 (04) : 427 - 441
  • [28] Adaptive statistical scheduling of divisible workloads in heterogeneous systems
    Horacio González-Vélez
    Murray Cole
    Journal of Scheduling, 2010, 13 : 427 - 441
  • [29] Co-scheduling HPC workloads on cache-partitioned CMP platforms
    Aupy, Guillaume
    Benoit, Anne
    Goglin, Brice
    Pottier, Loic
    Robert, Yves
    INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS, 2019, 33 (06): : 1221 - 1239
  • [30] Co-scheduling HPC workloads on cache-partitioned CMP platforms
    Aupy, Guillaume
    Benoit, Anne
    Goglin, Brice
    Pottier, Loic
    Robert, Yves
    2018 IEEE INTERNATIONAL CONFERENCE ON CLUSTER COMPUTING (CLUSTER), 2018, : 348 - 358