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. 发明专利

JP2006120182A Method for calculating parallel efficiency 有权
标题翻译：计算并行效率的方法
公开(公告)号：JP2006120182A
公开(公告)日：2006-05-11
申请号：JP2006001754
申请日：2006-01-06
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO
IPC分类号： G06F11/34
摘要： PROBLEM TO BE SOLVED: To apply a method for calculating parallel efficiency to many parallel processings including hetero processor environment even when load balance is not maintained and to perform quantitative association of parallel efficiency, load balance contribution ratio, virtual parallelization ratio which are parallel efficiency and parallel performance evaluation indices and a parallel performance obstruction factor. SOLUTION: The parallel efficiency E p (p) is calculated as follows using the load balance contribution ratio Rb(p) representing balance degrees of loads between each processor included in a parallel computer system, the virtual parallelization ratio Rp(p) representing ratio of parts calculated in parallel by each processor of processings executed in the parallel computer system for time and parallel performance obstruction factor contribution ratio Rj(p) representing ratio of processing time of each parallel performance obstruction factor part to all the processing times of all the processors included in the parallel computer system. COPYRIGHT: (C)2006,JPO&NCIPI
摘要翻译：要解决的问题：即使在不保持负载平衡的情况下，也应用并行计算并行处理的方法，包括异或处理器环境，并且执行并行效率，负载均衡贡献率，虚拟并行化比率并行效率和并行性能评估指标以及并行性能障碍因素。解决方案：使用负载平衡贡献率Rb（p）计算平行效率E p （p），负载平衡贡献率Rb（p）表示并行计算机系统中包括的每个处理器之间的负载平衡度，虚拟并行化比率Rp（p）表示在并行计算机系统中执行的处理的每个处理器并行计算的部分的时间和并行性能阻塞因子贡献率Rj（p）的比率表示每个并行性能障碍因子部分的处理时间的比率涉及并行计算机系统中包括的所有处理器的所有处理时间。版权所有（C）2006，JPO＆NCIPI

2. 发明专利

JP2004341750A Parallel performance estimation method and program 有权
公开(公告)号：JP2004341750A
公开(公告)日：2004-12-02
申请号：JP2003136491
申请日：2003-05-14
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO
IPC分类号： G06F11/34 , G06F15/177
摘要： PROBLEM TO BE SOLVED: To calculate an estimation index of parallel performance such as parallel efficiency of a parallel computer system by one-time measurement, without requiring multiple-time measurement performed while changing the number of processors. SOLUTION: By newly defining expression (6) as the parallel efficiency, the parallel efficiency can be calculated from only data obtainable by the one-time measurement without requiring a value such as τ(1). In the expression (6), p is the number of the processors; τ(p) is a processing time when it is executed by the processors of a number p; and γ i (p) is a processing time of a portion wherein parallel calculation is performed by the i-th processor (wherein, 1≤i≤p, and p is the number of the processors). COPYRIGHT: (C)2005,JPO&NCIPI

3. 发明专利

JP2004054680A Parallel efficiency calculation method 审中-公开
公开(公告)号：JP2004054680A
公开(公告)日：2004-02-19
申请号：JP2002212387
申请日：2002-07-22
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO
IPC分类号： G06F5/00 , G06F9/50 , G06F11/34 , G06F15/177 , G06F17/10
CPC分类号： G06F11/3433 , G06F11/3404 , G06F11/3419
摘要： PROBLEM TO BE SOLVED: To handle many types of parallel processing including a hetero processor environment even if load balance is not ensured, and quantitatively relate parallel efficiency to a load balance contribution ratio and a virtual parallelism ratio that are each a parallel performance evaluation index and to a parallel performance inhibiting factor. SOLUTION: Parallel efficiency E p (p) is calculated according to Formula 1 using a load balance contribution ratio Rb(p) indicating a load equilibrium degree among processors included in a parallel computer system, a virtual parallelism ratio Rp(p) indicating a time ratio, to processing executed in the parallel computer system, of a part of parallel calculation by each processor, and a parallel performance inhibiting factor contribution ratio Rj(p) indicating a ratio, to a total processing time of all the processors included in the parallel computer system, of a processing time of each parallel performance inhibiting factor part. COPYRIGHT: (C)2004,JPO

4. 发明专利

JP2013175065A Performance evaluation method, information processing device and program 有权
标题翻译：性能评估方法，信息处理设备和程序
公开(公告)号：JP2013175065A
公开(公告)日：2013-09-05
申请号：JP2012039406
申请日：2012-02-24
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ANAI HIROKAZU , ORII SHIGEO
IPC分类号： G06F11/34
摘要： PROBLEM TO BE SOLVED: To evaluate accuracy of a prediction of calculation time required by a computer system.SOLUTION: A storage unit 11 stores measurement information 11a indicating multiple calculation times measured by causing a computer system 20 to execute a sample program, with resource quantity changed. On the basis of the measurement information 11a, an arithmetic unit 12 determines a value of a parameter used for an evaluation model to calculate the relation between the resource quantity and the calculation time. The arithmetic unit 12 calculates variation of the value of the parameter when the difference between the calculation times indicated by the measurement information 11a and calculation time calculated according to the evaluation model is allowed to be widened to a threshold value. The arithmetic unit 12 outputs information indicating influence of the variation of the value of the parameter on the relation between the resource quantity and the calculation time.
摘要翻译：要解决的问题：评估计算机系统所需的计算时间的预测的准确性。解决方案：存储单元11存储指示通过使计算机系统20执行样本程序所测量的多个计算时间的测量信息11a，资源量改变了基于测量信息11a，算术单元12确定用于评估模型的参数的值，以计算资源量与计算时间之间的关系。当允许将由测量信息11a指示的计算时间与根据评估模型计算的计算时间之间的差扩大到阈值时，运算单元12计算参数值的变化。算术单元12输出指示参数值的变化对资源量与计算时间之间的关系的影响的信息。

5. 发明专利

JP2010033616A Model parameter calculating precision determination program, and storage medium storing the program 有权
标题翻译：模型参数计算精度确定程序和存储程序存储介质
公开(公告)号：JP2010033616A
公开(公告)日：2010-02-12
申请号：JP2009260262
申请日：2009-11-13
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO , ANAI HIROKAZU
IPC分类号： G06F19/00 , G06F19/12
摘要： PROBLEM TO BE SOLVED: To determine the accuracy of model parameter calculation by deciding the range of the numerical value of a model parameter using a quantifier elimination method.
SOLUTION: A program is configured to make a computer execute a procedure 1 for reading a model from a storage device; a procedure 2 for calculating a differential formula by replacing a variable with time series numerical values stored in the storage device; a procedure 3 for determining the range of the parameter of a model by using a quantifier elimination method; and a procedure 4 for determining the digit number of number showing the accuracy of calculation from the determined range of the model parameter.
COPYRIGHT: (C)2010,JPO&INPIT
摘要翻译：要解决的问题：通过使用量化器消除方法确定模型参数的数值范围来确定模型参数计算的准确性。解决方案：程序被配置为使计算机执行从存储设备读取模型的过程1; 用于通过用存储在存储装置中的时间序列数值替换变量来计算微分公式的过程2; 用于通过使用量化器消除方法来确定模型的参数的范围的过程3; 以及从确定的模型参数的范围确定表示计算精度的数字的数字编号的程序4。版权所有（C）2010，JPO＆INPIT

6. 发明专利

JP2010033615A Program and apparatus for determining model parameter in simulation 审中-公开
标题翻译：用于确定模拟参数的程序和装置
公开(公告)号：JP2010033615A
公开(公告)日：2010-02-12
申请号：JP2009260218
申请日：2009-11-13
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO , ANAI HIROKAZU
IPC分类号： G06F19/00
摘要： PROBLEM TO BE SOLVED: To efficiently and effectively determine the parameter of a model in simulation. SOLUTION: The model parameter determination program is configured to make a computer execute steps of: receiving input of data on a variable in a model of a system to be analyzed; generating a constraint expression relating to a parameter of a model based on the received data and the model of the system; and calculating a solution and/or a possible range of a feasible parameter by applying a quantifier elimination algorithm to the constraint expression. COPYRIGHT: (C)2010,JPO&INPIT
摘要翻译：要解决的问题：有效和有效地确定模拟中的模型参数。解决方案：模型参数确定程序被配置为使计算机执行以下步骤：接收关于要分析的系统的模型中的变量的数据的输入; 基于接收到的数据和系统的模型生成与模型的参数相关的约束表达式; 以及通过对所述约束表达式应用量化器消除算法来计算可行参数的解和/或可能范围。版权所有（C）2010，JPO＆INPIT

7. 发明专利

JP2007272478A Model parameter determination program and determination device 审中-公开
标题翻译：模型参数确定程序和确定设备
公开(公告)号：JP2007272478A
公开(公告)日：2007-10-18
申请号：JP2006096046
申请日：2006-03-30
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO , ANAI HIROKAZU
IPC分类号： G06F19/00
CPC分类号： G06F17/18 , G06F19/12
摘要： PROBLEM TO BE SOLVED: To execute, for example, the operation of an error upper limit value as a maximum value of an error term corresponding to a variable of a model within a practical time with a memory capacity for use. SOLUTION: A model parameter determination program causes a computer to execute; a procedure of substituting variables, variables calculated by initial values of model parameters, and differential values of variables into a first-order predicate logic formula; a procedures of substituting assumes values into unknown parameters in the logic formula after the substitution; and a procedure of applying a quantifier elimination method to the formula to obtain the error upper limit value. COPYRIGHT: (C)2008,JPO&INPIT
摘要翻译：要解决的问题：例如，在具有使用存储容量的实际时间内，执行误差上限值作为与模型的变量相对应的误差项的最大值的操作。
解决方案：模型参数确定程序使计算机执行; 将变量，模型参数的初始值计算的变量和变量的微分值替换为一阶谓词逻辑公式的过程; 替代后将逻辑公式中的假设值代入未知参数的程序; 以及对公式应用量化器消除方法以获得误差上限值的过程。版权所有（C）2008，JPO＆INPIT

8. 发明专利

JP2003050721A Method and device for parallel efficiency calculation 审中-公开
标题翻译：用于并行效率计算的方法和设备
公开(公告)号：JP2003050721A
公开(公告)日：2003-02-21
申请号：JP2001241121
申请日：2001-08-08
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO
IPC分类号： G06F11/34 , G06F15/16
CPC分类号： G06F11/3404 , G06F11/3419 , G06F11/3447 , G06F2201/86 , G06F2201/88
摘要： PROBLEM TO BE SOLVED: To clear causes to inhibit performance by quantitatively coupling the value of parallel efficiency with a factor for inhibiting improvement of the performance.
SOLUTION: Processing time α(p,n) of successive processing parts, processing time β(p,n)/p of a parallel processing part and processing time σ(p,n) to be cased by overhead of parallel processings, in the case of execution of a parallel processing programs, are acquired and a pararellization rate R
para (p,n), a successive calculation time ratio R
α (p,n), a parallel overhead ratio R
σ are calculated by using an acquired processing time α(p,n) for the successive processing part, a processing time β(p,n)/p for the parallel processing part and processing time σ(p,n) to be cased by overhead of parallel processings. Then, the parallel efficiency E
para (p,n) is calculated by 1/R
para (p,n)×(1-R
α (p,n)-R
σ (p,n)).
COPYRIGHT: (C)2003,JPO
摘要翻译：要解决的问题：通过将并行效率的值定量耦合以抑制性能提高的因素来清除抑制性能的原因。解决方案：平行处理部分的处理时间α（p，n），并行处理部分的处理时间β（p，n）/ p和由并行处理的开销来处理的处理时间σ（p，n）获取并行处理程序的执行情况，并且通过使用获取的处理时间α（p）来计算平行开销比Rα（p，n），连续计算时间比Rα（p，n），n），对于并行处理部分的处理时间β（p，n）/ p和由并行处理的开销来处理的处理时间σ（p，n）。然后，通过1 / Rpara（p，n）×（1-Rα（p，n）-Rσ（p，n））计算并行效率Epara（p，n）

9. 发明专利

JP2006106817A Model parameter computing accuracy judgement program and storage medium for storing the program 审中-公开
标题翻译：模型参数计算准确性评估程序和存储程序的存储介质
公开(公告)号：JP2006106817A
公开(公告)日：2006-04-20
申请号：JP2004288404
申请日：2004-09-30
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO , ANAI HIROKAZU
IPC分类号： G06F19/00
CPC分类号： G06F19/12
摘要： PROBLEM TO BE SOLVED: To judge the computing accuracy of a model parameter by determining the range of the numerics of the model parameter by using a limited symbol elimination method. SOLUTION: This program is provided to make a computer execute a procedure 1 for reading a model from a storage device, a procedure 2 for calculating a differential formula by replacing variables with time series numerics stored in the storage device, a procedure 3 for determining the range of the parameter of the model by using a limited symbol elimination method and a procedure 4 for judging the number of digits of effective numbers showing computing accuracy from the determined range of the model parameter. COPYRIGHT: (C)2006,JPO&NCIPI
摘要翻译：要解决的问题：通过使用有限符号消除方法确定模型参数的数值范围来判断模型参数的计算精度。解决方案：提供该程序以使计算机执行从存储设备读取模型的过程1，用于通过用存储在存储设备中的时间序列数字替换变量来计算差分公式的过程2，过程3 用于通过使用有限符号消除方法来确定模型的参数的范围;以及用于从所确定的模型参数的范围来判断显示计算精度的有效数字的位数的过程4。版权所有（C）2006，JPO＆NCIPI

10. 发明专利

JP2005129015A Program and apparatus for determining model parameter in simulation 审中-公开
标题翻译：用于确定模拟参数的程序和装置
公开(公告)号：JP2005129015A
公开(公告)日：2005-05-19
申请号：JP2004254903
申请日：2004-09-01
申请人： Fujitsu Ltd , 富士通株式会社
发明人： ORII SHIGEO , ANAI HIROKAZU
IPC分类号： G06F17/10 , G06F17/50 , G06F19/00 , G06F19/12
CPC分类号： G06F17/50 , G06F2217/06
摘要： PROBLEM TO BE SOLVED: To efficiently and effectively determine model parameters in simulations.
SOLUTION: With model parameter determining programs, it is structured so as to allow a computer to execute a step for receiving input of parameter data in a model for a system that is an analysis target, a step for generating restriction equations related to parameters in the model based on the received data and the model for the system, and a step for acquiring solution existence and/or an existence range of executable parameters by applying a quantifier elimination algorithm to the restriction equations.
COPYRIGHT: (C)2005,JPO&NCIPI

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