专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明授权

US09794138B2 Computer system, method, and program 有权
公开(公告)号：US09794138B2
公开(公告)日：2017-10-17
申请号：US13695558
申请日：2011-04-21
申请人： Munehiro Doi , Hideaki Komatsu , Kumiko Maeda , Masana Murase , Takeo Yoshizawa
发明人： Munehiro Doi , Hideaki Komatsu , Kumiko Maeda , Masana Murase , Takeo Yoshizawa
IPC分类号： H04L12/26 , G06F9/48
CPC分类号： H04L43/04 , G06F9/485
摘要： Traffic data while the system is in operation is collected for a certain time as a preprocess. Typical patterns are extracted from the collected traffic data. Next, stream programs are created for the individual typical patterns and stored for the future reference. Next, the IDs of alternative tasks for transition among different stream programs are stored. In actual system operation, the system measures traffic data regularly or at any time, compares the resultant patterns with the typical patterns, and selects a stream program corresponding to the closest typical pattern as the next phase. Program shutdown time when shifting from the stream program in the present phase to the next phase can be reduced by gradually shifting empty tasks in the present phase to the next stream program as alternative tasks in consideration of the cost of switching between tasks, the cost of transferring data among resources, and so on.

2. 发明申请

US20110131554A1 APPLICATION GENERATION SYSTEM, METHOD, AND PROGRAM PRODUCT 审中-公开
标题翻译：应用生成系统，方法和程序产品
公开(公告)号：US20110131554A1
公开(公告)日：2011-06-02
申请号：US12955147
申请日：2010-11-29
申请人： Munehiro Doi , Hideaki Komatsu , Kumiko Maeda , Masana Murase , Takeo Yoshizawa
发明人： Munehiro Doi , Hideaki Komatsu , Kumiko Maeda , Masana Murase , Takeo Yoshizawa
IPC分类号： G06F9/44
CPC分类号： G06F8/4441
摘要： A method, system and computer program product for optimizing performance of an application running on a hybrid system. The method includes the steps of: selecting a first user defined operator from a library component within the application; determining at least one available hardware resource; generating at least one execution pattern for the first user defined operator based on the available hardware resource; compiling the execution pattern; measuring the execution speed of the execution pattern on the available hardware resource; and storing the execution speed and the execution pattern in an optimization table; where at least one of the steps is carried out using a computer device so that performance of said application is optimized on the hybrid system.
摘要翻译：一种用于优化在混合系统上运行的应用的性能的方法，系统和计算机程序产品。该方法包括以下步骤：从应用程序内的库组件中选择第一用户定义的运算符; 确定至少一个可用硬件资源; 基于所述可用硬件资源为所述第一用户定义的运算符生成至少一个执行模式; 编译执行模式; 测量可用硬件资源上执行模式的执行速度; 并将执行速度和执行模式存储在优化表中; 其中使用计算机设备执行至少一个步骤，使得在混合系统上优化所述应用的性能。

3. 发明申请

US20130103829A1 COMPUTER SYSTEM, METHOD, AND PROGRAM 有权
标题翻译：计算机系统，方法和程序
公开(公告)号：US20130103829A1
公开(公告)日：2013-04-25
申请号：US13695558
申请日：2011-04-21
申请人： Munehiro Doi , Hideaki Komatsu , Kumiko Maeda , Masana Murase , Takeo Yoshizawa
发明人： Munehiro Doi , Hideaki Komatsu , Kumiko Maeda , Masana Murase , Takeo Yoshizawa
IPC分类号： H04L12/26
CPC分类号： H04L43/04 , G06F9/485
摘要： Traffic data while the system is in operation is collected for a certain time as a preprocess. Typical patterns are extracted from the collected traffic data. Next, stream programs are created for the individual typical patterns and stored for the future reference. Next, the IDs of alternative tasks for transition among different stream programs are stored. In actual system operation, the system measures traffic data regularly or at any time, compares the resultant patterns with the typical patterns, and selects a stream program corresponding to the closest typical pattern as the next phase. Program shutdown time when shifting from the stream program in the present phase to the next phase can be reduced by gradually shifting empty tasks in the present phase to the next stream program as alternative tasks in consideration of the cost of switching between tasks, the cost of transferring data among resources, and so on.
摘要翻译：系统运行时的流量数据作为预处理收集一段时间。从收集的交通数据中提取典型模式。接下来，为各个典型模式创建流程序，并存储以供将来参考。接下来，存储用于不同流程序之间的转换的替代任务的ID。在实际系统运行中，系统定期或随时测量交通数据，将结果模式与典型模式进行比较，并选择与最接近的典型模式对应的流程序作为下一阶段。考虑到任务之间的切换成本，可以通过逐渐将当前阶段的空任务逐步转移到下一个流程序来将程序关闭时间从当前阶段的流程序转换到下一阶段，成本资源之间传输数据等等。

4. 发明申请

US20100138810A1 PARALLELING PROCESSING METHOD, SYSTEM AND PROGRAM 有权
标题翻译：并行处理方法，系统与程序
公开(公告)号：US20100138810A1
公开(公告)日：2010-06-03
申请号：US12629114
申请日：2009-12-02
申请人： Hideaki Komatsu , Arquimedes Martinez Canedo , Takeo Yoshizawa
发明人： Hideaki Komatsu , Arquimedes Martinez Canedo , Takeo Yoshizawa
IPC分类号： G06F9/44 , G06F9/38
CPC分类号： G06F8/30 , G06F8/456
摘要： Paralleling processing system and method. When clusters are formed based on strongly connected components, a single cluster (fat cluster) having at least a predetermined number of blocks, or an expected processing time exceeding a predetermined threshold, is formed. The fat cluster is subjected to an unrolling process to make multiple copies of the processing of the fat cluster and to assign the copies to individual processors. Processing of the fat cluster is executed by the multiple processor devices in a pipelined manner. If a fat cluster to be iteratively executed cannot be executed in the pipelined manner because a processing result of an nth iteration of the fat cluster depends on a processing result of a preceding iteration of the fat cluster an input value needed for execution of the fat cluster is generated based on a certain prediction, and the fat cluster is speculatively executed.
摘要翻译：并行处理系统及方法。当基于强连接的组件形成集群时，形成具有至少预定数量块的单个集群（胖集群）或超过预定阈值的预期处理时间。脂肪群体经历展开过程，以便对脂肪群集的处理进行多个拷贝，并将副本分配给各个处理器。脂肪簇的处理由多个处理器设备以流水线方式执行。如果要迭代执行的胖群集不能以流水线方式执行，因为胖群的第n次迭代的处理结果取决于胖群的前一迭代的处理结果，则执行胖群所需的输入值是基于某种预测生成的，并且推测性地执行脂肪群。

5. 发明授权

US08407679B2 Source code processing method, system and program 有权
标题翻译：源代码处理方法，系统和程序
公开(公告)号：US08407679B2
公开(公告)日：2013-03-26
申请号：US12603598
申请日：2009-10-22
申请人： Hideaki Komatsu , Takeo Yoshizawa
发明人： Hideaki Komatsu , Takeo Yoshizawa
IPC分类号： G06F9/45
CPC分类号： G06F8/70 , G06F8/45 , G06F8/456
摘要： A method, system, and computer readable article of manufacture to enable parallel execution of a divided source code in a multiprocessor system. The method includes the steps of: inputting an original source code by an input device into the computing apparatus; finding a critical path in the original source code by a critical path cut module; cutting the critical path in the original source code into a plurality of process block groups by the critical path cut module; and dividing the plurality of process block groups among a plurality of processors in the multiprocessor system by a CPU assignment code generation module to produce the divided source code. The system includes an input device; a critical path cut module; and a CPU assignment code generation unit to produce the divided source code. The computer readable article of manufacture includes instructions to implement the method.
摘要翻译：一种方法，系统和计算机可读制品，用于在多处理器系统中并行执行分割的源代码。该方法包括以下步骤：通过输入装置将原始源代码输入到计算装置中; 通过关键路径切割模块找到原始源代码中的关键路径; 通过关键路径切割模块将原始源代码中的关键路径切割成多个进程块组; 以及通过CPU分配码产生模块在多处理器系统中的多个处理器之间划分多个处理块组，以产生分割的源代码。该系统包括输入装置; 关键路径切割模块; 以及CPU分配代码生成单元，用于产生分割的源代码。计算机可读制品包括实现该方法的指令。

6. 发明申请

US20110107162A1 PARALLELIZATION METHOD, SYSTEM AND PROGRAM 失效
标题翻译：并行化方法，系统与程序
公开(公告)号：US20110107162A1
公开(公告)日：2011-05-05
申请号：US12913822
申请日：2010-10-28
申请人： Arquimedes Martinez Canedo , Hideaki Komatsu , Takeo Yoshizawa
发明人： Arquimedes Martinez Canedo , Hideaki Komatsu , Takeo Yoshizawa
IPC分类号： G06F11/26
CPC分类号： G06F11/261
摘要： A computer-implemented method, system, and article of manufacture for parallelizing a code configured by coupling a functional block having an internal state and a functional block without any internal state. The method includes: creating and storing a graphical representation where functional blocks are chosen as nodes and connections between functional blocks are chosen as links; visiting the nodes on the graphical representation sequentially, detecting inputs from functional blocks without any internal state to functional blocks having an internal state and storing these functional blocks as a set of use blocks, and detecting inputs from functional blocks having an internal state to functional blocks without any internal state and storing these functional blocks as a set of definition blocks; and forming strands of functional blocks based on information on the set of use blocks and information on the set of definition blocks stored in association with the functional blocks.
摘要翻译：一种用于并行化代码的计算机实现的方法，系统和制品，其通过耦合具有内部状态的功能块和功能块而没有任何内部状态来配置。该方法包括：创建和存储其中功能块被选择为节点的图形表示，并且功能块之间的连接被选择为链接; 在图形表示上顺序地访问节点，检测来自功能块的输入，而没有任何内部状态到具有内部状态的功能块，并将这些功能块存储为一组使用块，并且检测具有内部状态的功能块到功能块的输入没有任何内部状态并将这些功能块存储为一组定义块; 以及基于关于所述使用块集合的信息和与所述功能块相关联地存储的所述定义块集合上的信息来形成功能块组。

7. 发明申请

US20110083125A1 PARALLELIZATION PROCESSING METHOD, SYSTEM AND PROGRAM 审中-公开
标题翻译：并行处理方法，系统与程序
公开(公告)号：US20110083125A1
公开(公告)日：2011-04-07
申请号：US12898851
申请日：2010-10-06
申请人： Hideaki Komatsu , Takeo Yoshizawa
发明人： Hideaki Komatsu , Takeo Yoshizawa
IPC分类号： G06F9/45
CPC分类号： G06F8/456
摘要： A unified parallelization table is formed by describing a process, to be executed, with a plurality of control blocks and edges connecting the control blocks; selecting highly predictable edges from the edges; identifying strongly-connected clusters; creating a parallelization table, having the entries of the number of processors, the costs thereof and corresponding clusters, for each node in the strongly-connected clusters and a non-strongly connected cluster between the strongly-connected clusters; creating a graph consisting of parallelization tables; converting the graph consisting of the parallelization tables into a series-parallel graph; and merging the parallelization tables for each serial path merging the parallelization tables for each parallel section. Then, based on the number of processors and the cost value in the unified parallelization table, a best entry is selected and an executable code to be allocated to each processor is generated.
摘要翻译：通过用连接控制块的多个控制块和边缘描述要执行的处理来形成统一的并行表; 从边缘选择高度可预测的边缘; 识别强连接的群集; 创建一个并行化表，具有针对强连接集群中的每个节点的处理器数量，成本和对应集群的条目以及强连接集群之间的非强连接集群; 创建一个由并行表组成的图; 将由并行化表组成的图转换成串并行图; 并且合并每个并行段的并行表的每个串行路径的并行化表。然后，基于处理器的数量和统一并行化表中的成本值，选择最佳条目，并生成要分配给每个处理器的可执行代码。

8. 发明授权

US08438553B2 Paralleling processing method, system and program 有权
标题翻译：并行处理方法，系统和程序
公开(公告)号：US08438553B2
公开(公告)日：2013-05-07
申请号：US12629114
申请日：2009-12-02
申请人： Hideaki Komatsu , Arquimedes Martinez Canedo , Takeo Yoshizawa
发明人： Hideaki Komatsu , Arquimedes Martinez Canedo , Takeo Yoshizawa
IPC分类号： G06F9/45
CPC分类号： G06F8/30 , G06F8/456
摘要： Paralleling processing system and method. When clusters are formed based on strongly connected components, a single cluster (fat cluster) having at least a predetermined number of blocks, or an expected processing time exceeding a predetermined threshold, is formed. The fat cluster is subjected to an unrolling process to make multiple copies of the processing of the fat cluster and to assign the copies to individual processors. Processing of the fat cluster is executed by the multiple processor devices in a pipelined manner. If a fat cluster to be iteratively executed cannot be executed in the pipelined manner because a processing result of an nth iteration of the fat cluster depends on a processing result of a preceding iteration of the fat cluster an input value needed for execution of the fat cluster is generated based on a certain prediction, and the fat cluster is speculatively executed.
摘要翻译：并行处理系统及方法。当基于强连接的组件形成集群时，形成具有至少预定数量块的单个集群（胖集群）或超过预定阈值的预期处理时间。脂肪群体经历展开过程，以便对脂肪群集的处理进行多个拷贝，并将副本分配给各个处理器。脂肪簇的处理由多个处理器设备以流水线方式执行。如果要迭代执行的胖群集不能以流水线方式执行，因为胖群的第n次迭代的处理结果取决于胖群的前一迭代的处理结果，则执行胖群所需的输入值是基于某种预测生成的，并且推测性地执行脂肪群。

9. 发明授权

US08959498B2 Parallelization method, system and program 有权
标题翻译：并行化方法，系统和程序
公开(公告)号：US08959498B2
公开(公告)日：2015-02-17
申请号：US13031666
申请日：2011-02-22
申请人： Hideaki Komatsu , Takeo Yoshizawa
发明人： Hideaki Komatsu , Takeo Yoshizawa
IPC分类号： G06F9/45
CPC分类号： G06F8/456
摘要： A parallelization method, system and program. A program expressed by a block diagram or the like is divided into strands and a balance in calculation time is made among the strands. The functional blocks are divided into strands and the strand involving the maximum calculation time from a strand set is found. One or more movable blocks in the strand involving the maximum calculation time is found. The next step is obtaining calculation time of each strand after the movable block is moved to the strand in the input or output direction according to its property, and moving the block to a strand most largely reducing the calculation time of the strand having the maximum calculation time before the movement. This process loops until calculation time is no longer reduced. Strands are then transformed into source codes. Source codes are compiled and assigned to separate cores or processors for execution.
摘要翻译：并行化方法，系统和程序。通过框图等表示的程序被划分成多个线，并且在线之间进行计算时间的平衡。功能块被划分为股线，并且发现涉及从股线组得到的最大计算时间的股线。发现涉及最大计算时间的股份中的一个或多个可移动块。下一步是根据其可移动块在输入或输出方向上移动到股线之后获得每条股的计算时间，并且将块移动到股线，最大程度上减少具有最大计算的股线的计算时间运动前的时间。此过程循环，直到计算时间不再减少。然后将链条转换为源代码。源代码被编译并分配给单独的核心或处理器以供执行。

10. 发明授权

US08677334B2 Parallelization method, system and program 失效
标题翻译：并行化方法，系统和程序
公开(公告)号：US08677334B2
公开(公告)日：2014-03-18
申请号：US12913822
申请日：2010-10-28
申请人： Arquimedes Martinez Canedo , Hideaki Komatsu , Takeo Yoshizawa
发明人： Arquimedes Martinez Canedo , Hideaki Komatsu , Takeo Yoshizawa
IPC分类号： G06F9/45
CPC分类号： G06F11/261
摘要： A computer-implemented method, system, and article of manufacture for parallelizing a code configured by coupling a functional block having an internal state and a functional block without any internal state. The method includes: creating and storing a graphical representation where functional blocks are chosen as nodes and connections between functional blocks are chosen as links; visiting the nodes on the graphical representation sequentially, detecting inputs from functional blocks without any internal state to functional blocks having an internal state and storing these functional blocks as a set of use blocks, and detecting inputs from functional blocks having an internal state to functional blocks without any internal state and storing these functional blocks as a set of definition blocks; and forming strands of functional blocks based on information on the set of use blocks and information on the set of definition blocks stored in association with the functional blocks.
摘要翻译：一种用于并行化代码的计算机实现的方法，系统和制品，其通过耦合具有内部状态的功能块和功能块而没有任何内部状态来配置。该方法包括：创建和存储其中功能块被选择为节点的图形表示，并且功能块之间的连接被选择为链接; 在图形表示上顺序地访问节点，检测来自功能块的输入，而没有任何内部状态到具有内部状态的功能块，并将这些功能块存储为一组使用块，并且检测具有内部状态的功能块到功能块的输入没有任何内部状态并将这些功能块存储为一组定义块; 以及基于关于所述使用块集合的信息和与所述功能块相关联地存储的所述定义块集合上的信息来形成功能块组。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式