CN101872304A - Clustered operation system technology based method for improving scalability of many-core application program - Google Patents
Clustered operation system technology based method for improving scalability of many-core application program Download PDFInfo
- Publication number
- CN101872304A CN101872304A CN201010197968A CN201010197968A CN101872304A CN 101872304 A CN101872304 A CN 101872304A CN 201010197968 A CN201010197968 A CN 201010197968A CN 201010197968 A CN201010197968 A CN 201010197968A CN 101872304 A CN101872304 A CN 101872304A
- Authority
- CN
- China
- Prior art keywords
- operating system
- scalability
- many
- application program
- operation system
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention belongs to the technical field of computer system software, particularly relates to a clustered operation system technology based method for improving the scalability of a many-core application program. The invention provides high scalability support for a superior many-core application program by using a clustered operation system technology. A clustered operation system comprises a virtual machine monitor with high scalability, a remote process collaboration module, a resource synchronization module and a system service virtual layer. The invention is based on the main point of operating a plurality of general operation systems in a single-machine many-core hardware environment and providing system service for a single many-core application program at the same time so as to avoid the limit to superior application on scalability at the level of operation system. The invention effectively improves the scalability of an application program in a many-core environment.
Description
Technical field
The invention belongs to the computer system software technical field, be specifically related to the method for a kind of raising towards the scalability of the application program of many nuclear environments.
Background technology
Along with energy consumption and the heat dissipation problem of CPU are apparent more outstanding, promoting CPU frequency merely no longer has been a kind of method that increases system performance cost-effectively.Therefore, as the continuity of Moore's Law, multinuclear (multi-core) and even many nuclear (many-core) technology are popularized in industrial community gradually.Many nuclear technology continue to promote the computing power of processor by the quantity that constantly increases the calculating core, thereby continue " Moore's Law ".Arise at the historic moment towards the application program of many nuclear environments in a large number, they originally have good scalability in application layer, can give full play to many nuclear environment performances.But do not increase along with the increase of processor core number towards the performance of the application program of many nuclear environments is simple thereupon.This mainly is because the application program on upper strata is subjected to the restriction from operating system layer.Therefore, examine the today that has become main flow processor technology the crowd, existing operating system need be transformed, optimize and redesign at many nuclear environments, thereby improves the scalabilities of many nuclear application program integral body.
The appearance of many nuclear technology has brought new demand for the exploitation of operating system, and the ratio that application program was carried out in operating system in the execution time constantly increases.Yet unfortunately, existing daily operating system is all come from the monokaryon era development mostly.A lot of modules do not fully take into account the support to many nuclear system structures in design originally.Exist the system services of a large amount of restriction upper level applications scalabilities to realize in the operating system, share the management of cache resources such as memory-mapped.These services show good performance under the monokaryon environment, but under many nuclear environments because for the aggravation of system resource competition, performance worsens even occurs falling back thereupon.
In the face of operating system can't provide efficient scalability support issue under many nuclear environments, there are two kinds of solutions at present.Short-term strategies is that existing daily operating system is carried out local optimum, such as the local runtime formation of RCU data structure that adopts fine granularity lock optimisation technique and the shared competition technology of elimination.But because the high complicacy and the coupling of operating system itself, and the compatible requirement that must satisfy, the optimisation strategy of this locality often be difficult to realization and effect limited.For example Windows7 operating system just relates to surpassing the modification of 6000 line codes in 58 files for the modification of distributor spin lock.Long term policy on the other side is taked more radical means, finishes the operating system of developing again towards many nuclear environments.Provide efficient support by adopting in a large number to the upper level applications scalability at the realization criterion of many nuclear environments and optimisation strategy.Such example has the Corey operating system by common exploitation such as masschusetts, u.s.a Polytechnics and Fudan University, has by the Barrelfish operating system of the common exploitation of Zurich, SUI ETH university and Cambridge research institute of Microsoft etc.But these operating systems all face the problem with original daily Compatibility of Operating System, and to examine the support of application program towards the crowd also very limited to existing, and for the consideration of marketing factor, want to replace very difficulty of existing daily operating system.
Because application program depends on the service that operating system provides, the competition for system resource under many nuclear environments becomes the main bottleneck that suppresses many nuclear application performances.
Summary of the invention
The object of the present invention is to provide a kind of raising to examine the method for the scalability of application program,, when keeping Compatibility of Operating System, improve scalability so that provide the support of the crowd being examined application program in the operating system aspect towards the crowd.
The method of the scalability of application program is examined in raising provided by the invention towards the crowd, its core concept is to move a plurality of general-purpose operating systems simultaneously under the many nuclear of unit hardware environment, be the service of single many nuclear application providing system simultaneously, thereby avoid upper layer application in the restriction aspect the scalability in the operating system aspect.Specifically, being is that upper strata many nuclear application program provides the high scalability support with the clustered operation system technology.
Described clustered operation system is mainly by the scalable monitor of virtual machine of height, remote process collaboration module, formations such as resource synchronization module and system service virtual level.High scalable virtual machine incorporates current in the state-of-the-art technology aspect the system scalability research on the basis of existing monitor of virtual machine.Utilize the relative operating system simple and flexible of monitor of virtual machine not have the characteristics of compatibility issue, for the clustered and the high scalability of operating system provides support.The remote process collaboration module is at process and the threading model realized between a plurality of operating systems under the single operating system.And resource synchronization module utilizes lower floor's monitor of virtual machine to realize that the resource scalability of cross operating system is shared.The three combines and has realized support to the collaborative work of a plurality of general-purpose operating system high scalabilities.The system service virtual level is loaded in the general-purpose operating system in the mode of operating system module, for upper level applications provides the DLL (dynamic link library) consistent with single operating system, is implemented in and does not lose compatibility when the high scalability support is provided.
The invention has the beneficial effects as follows, improve the scalability of application program under many nuclear environments.Utilize high scalable monitor of virtual machine to move a plurality of general-purpose operating systems and provide high scalability system service support, significantly improve application program capacity thereby make full use of many karyonide system resources for single application program.
Description of drawings
Fig. 1 is a clustered operation system technology configuration diagram of the present invention.
Embodiment
The present invention is further described below in conjunction with drawings and Examples.
Fig. 1 has described the framework of clustered operation system technology.Being positioned on many nuclear hardware environment is high scalable monitor of virtual machine, realizes support to the collaborative work of a plurality of general-purpose operating system high scalability with remote process collaboration module and resource synchronization module.Being positioned on the monitor of virtual machine is the general-purpose operating system of a plurality of unmodified, keeps the compatible and stable of system.Realized that as the system service virtual level that operating system module exists the conventional system services request of self-application layer is transmitted to a plurality of operating systems in the future, kept the compatibility of application programming interfaces.Total system can effectively improve the upper strata towards the scalability that the crowd examines application program, promotes application program capacity.
Claims (1)
1. one kind is improved the method for scalability of many-core application program based on clustered operation system technology, and it is characterized in that coming with clustered operation system technology is that upper strata many nuclear application programs provide the high scalability support; Wherein, described clustered operation system comprises: high scalable monitor of virtual machine, remote process collaboration module, resource synchronization module and system service virtual level; The relative operating system simple and flexible of the scalable monitor of virtual machine of described height does not have compatibility issue, for the clustered and the high scalability of operating system provides support; Described remote process collaboration module is at process and the threading model realized between a plurality of operating systems under the single operating system; Described resource synchronization module utilizes lower floor's monitor of virtual machine to realize that the resource scalability of cross operating system is shared; Above-mentioned three combines and realizes support to the collaborative work of a plurality of general-purpose operating system high scalabilities; Described system service virtual level is loaded in the general-purpose operating system in the mode of operating system module, for upper level applications provides the DLL (dynamic link library) consistent with single operating system, provides the high scalability support.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010197968A CN101872304A (en) | 2010-06-10 | 2010-06-10 | Clustered operation system technology based method for improving scalability of many-core application program |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010197968A CN101872304A (en) | 2010-06-10 | 2010-06-10 | Clustered operation system technology based method for improving scalability of many-core application program |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101872304A true CN101872304A (en) | 2010-10-27 |
Family
ID=42997178
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010197968A Pending CN101872304A (en) | 2010-06-10 | 2010-06-10 | Clustered operation system technology based method for improving scalability of many-core application program |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101872304A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156270A (en) * | 2014-07-28 | 2014-11-19 | 浪潮(北京)电子信息产业有限公司 | Method and system for single machine running multiple operating systems at the same time |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080077993A1 (en) * | 2006-09-26 | 2008-03-27 | Zimmer Vincent J | Methods and arrangements to launch trusted, co-existing environments |
| CN101169731A (en) * | 2007-12-05 | 2008-04-30 | 华为技术有限公司 | Multiple-path multiple-core server and its CPU virtualization processing method |
| CN101398769A (en) * | 2008-10-28 | 2009-04-01 | 北京航空航天大学 | Processor resource integrating and utilizing method transparent to operating system |
-
2010
- 2010-06-10 CN CN201010197968A patent/CN101872304A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080077993A1 (en) * | 2006-09-26 | 2008-03-27 | Zimmer Vincent J | Methods and arrangements to launch trusted, co-existing environments |
| CN101169731A (en) * | 2007-12-05 | 2008-04-30 | 华为技术有限公司 | Multiple-path multiple-core server and its CPU virtualization processing method |
| CN101398769A (en) * | 2008-10-28 | 2009-04-01 | 北京航空航天大学 | Processor resource integrating and utilizing method transparent to operating system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156270A (en) * | 2014-07-28 | 2014-11-19 | 浪潮(北京)电子信息产业有限公司 | Method and system for single machine running multiple operating systems at the same time |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Perumalla | /spl mu/sik-a micro-kernel for parallel/distributed simulation systems | |
| Dorier et al. | CALCioM: Mitigating I/O interference in HPC systems through cross-application coordination | |
| Schwarzkopf et al. | The seven deadly sins of cloud computing research | |
| Kaiser et al. | Hpx: A task based programming model in a global address space | |
| Kumar et al. | PAMI: A parallel active message interface for the Blue Gene/Q supercomputer | |
| Kirby et al. | An approach to ad hoc cloud computing | |
| Chen et al. | Tiled-MapReduce: optimizing resource usages of data-parallel applications on multicore with tiling | |
| Buyya et al. | Energy-efficient management of data center resources for cloud computing: a vision, architectural elements, and open challenges | |
| Porobic et al. | OLTP on hardware islands | |
| Anderson et al. | Treehouse: A case for carbon-aware datacenter software | |
| Zhou et al. | A case for software managed coherence in manycore processors | |
| CN106201720A (en) | Virtual symmetric multi-processors virtual machine creation method, data processing method and system | |
| CN103150217A (en) | Design method of multi-core processor operating system | |
| Ellsworth et al. | A unified platform for exploring power management strategies | |
| CN103092703A (en) | Synchronized method and device for Java resource access | |
| CN101872304A (en) | Clustered operation system technology based method for improving scalability of many-core application program | |
| Singh et al. | A survey on techniques to achive energy efficiency in cloud computing | |
| CN101739299A (en) | Method for dynamically and fairly partitioning shared cache based on chip multiprocessor | |
| Brady et al. | SmartGridRPC: The new RPC model for high performance Grid computing | |
| CN103294639A (en) | CPU+MIC mixed heterogeneous cluster system for achieving large-scale computing | |
| Chen et al. | Handling shared variable synchronization in multi-core Network-on-Chips with distributed memory | |
| Liu et al. | Virtual machine dynamic deployment scheme based on double-cursor mechanism | |
| Wang et al. | Fargraph+: Excavating the parallelism of graph processing workload on RDMA-based far memory system | |
| Papagiannis et al. | Scalable runtime support for data-intensive applications on the single-chip cloud computer | |
| Chai et al. | High-performance cloud simulation platform advanced research of cloud simulation platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20101027 |