Accelerating Parallel Applications in Cloud Platforms via Adaptive Time-Slice Control
- Huazhong Univ. of Science and Technology, Wuhan (China)
- Tencent Group, Guangdong (China)
- Alibaba Group, Zhejiang (China)
- Argonne National Lab. (ANL), Lemont, IL (United States)
Cloud platforms can provide flexible and cost-effective environments for parallel applications. However, the resource over-commitment issues, i.e., cloud providers often provide much more executable virtual CPUs than available physical CPUs, still impede the synchronization operations of parallel applications, causing severe performance degradation. Existing methods optimize parallel applications by promoting the priorities of involved VMs. They cannot fully explore the performance of parallel applications, because they ignore the time-slice requirements of different phases of parallel applications. Furthermore, non-parallel applications experience unsatisfied performance because of low scheduling priorities. Given empirical analysis on time-slices of virtual machines (VMs), we find that shortening time-slices can mitigate synchronization overhead which incurs during communication phases, while over-short time-slices cause frequent cache misses in computation phases. Accordingly, we propose an Adaptive Time-slice Control (ATC) mechanism. ATC first detects the phases of parallel applications based on lock latency or cache misses. Then, ATC shortens time-slices during communication phases and prolongs time-slices during computation phases for parallel applications, and sets a uniform time-slice for non-parallel applications. Finally, we evaluate ATC using seven well-known benchmarks with 25+ applications. Experiments show that ATC obtains 1.5-75x performance gain for running parallel applications than state-of-the-art solutions, with nearly unaffected impact on non-parallel applications.
- Research Organization:
- Argonne National Lab. (ANL), Argonne, IL (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC); National Key Research and Development Program of China; National Science Foundation of China
- Grant/Contract Number:
- AC02-06CH11357; 2018YFB1004805; 61872155; 61732010; 2019aea171
- OSTI ID:
- 1863258
- Journal Information:
- IEEE Transactions on Computers, Vol. 70, Issue 7; ISSN 0018-9340
- Publisher:
- IEEECopyright Statement
- Country of Publication:
- United States
- Language:
- English
Time-Sharing Parallel Applications with Performance Isolation and Control
|
conference | June 2007 |
Cloud versus in-house cluster: evaluating Amazon cluster compute instances for running MPI applications
|
conference | January 2011 |
Performance implications of virtualizing multicore cluster machines
|
conference | January 2008 |
Implicit coscheduling: coordinated scheduling with implicit information in distributed systems
|
journal | August 2001 |
The hybrid scheduling framework for virtual machine systems
|
conference | January 2009 |
Dynamic adaptive scheduling for virtual machines
|
conference | January 2011 |
A bridging model for parallel computation
|
journal | August 1990 |
Dynamic Acceleration of Parallel Applications in Cloud Platforms by Adaptive Time-Slice Control
|
conference | May 2016 |
vScale: automatic and efficient processor scaling for SMP virtual machines
|
conference | April 2016 |
Demand-based coordinated scheduling for SMP VMs
|
conference | January 2013 |
Synchronization-Aware Scheduling for Virtual Clusters in Cloud
|
journal | October 2015 |
The Scalasca performance toolset architecture
|
journal | January 2010 |
vSlicer: latency-aware virtual machine scheduling via differentiated-frequency CPU slicing
|
conference | January 2012 |
Xen and the art of virtualization
|
conference | January 2003 |
The impact of management operations on the virtualized datacenter
|
conference | January 2010 |
The NAS parallel benchmarks---summary and preliminary results
|
conference | January 1991 |
The PARSEC benchmark suite: characterization and architectural implications
|
conference | January 2008 |
Is co-scheduling too expensive for SMP VMs?
|
conference | January 2011 |
Towards fair and efficient SMP virtual machine scheduling
|
conference | January 2014 |
Dynamic Switching-Frequency Scaling: Scheduling Overcommitted Domains in Xen VMM
|
conference | September 2010 |
Flexible resource allocation for reliable virtual cluster computing systems
|
conference | January 2011 |
Threads vs. caches: Modeling the behavior of parallel workloads
|
conference | October 2010 |
Supporting Overcommitted Virtual Machines through Hardware Spin Detection
|
journal | February 2012 |
Characterizing and Optimizing the Performance of Multithreaded Programs Under Interference
|
conference | September 2016 |
Micro-Sliced Virtual Processors to Hide the Effect of Discontinuous CPU Availability for Consolidated Systems
|
conference | December 2014 |
Perfctr-Xen: a framework for performance counter virtualization
|
journal | July 2011 |
Similar Records
HUNTing the Overlap
Online anomaly detection for multiāsource VMware using a distributed streaming framework