Fujitsu Corporation

Tohru Matsumoto

Tadamitsu Ryu

Shingo Hirono

Toshio Takahara

Takanori Fukatsu

Yoshio Mogi

Takashi Ohshiro

Masahiko Murakawa

Gen Kakehi

Masao Tomita

Kazuo Watanabe

Noboru Asai

A method of visualizing the results of performance monitoring and analysis for a parallel computing system in which a plurality of processors execute a parallel processing program composed of a plurality of routines. First, information on the execution time of each routine is collected in a realtime manner while the program is concurrently executed by the plurality of processors. Second, a maximum, average, and minimum values of the execution time of each routine are calculated, based on the information collected for the plurality of processors. Third, the collected information is summarized as an execution profile and displayed in graphical form by using bar graphs, pie graphs, or radar charts. For each procedure or program loop, the present method clarifies the percentages of net user program execution, communication, synchronization, and other parallelization overheads, as well as indicating their deviations. This feature helps the user to understand the actual program behavior and to tune up the parallel processing program.

Method of visualizing results of performance monitoring and analysis in

A data retrieval method and apparatus in a distributed database system. Data is retrieved from a distributed database quickly with a minimum load on each terminal. A contents file is provided in each terminal, and a link file is provided in a center system to coordinate access to the distributed data in each terminal.

Distributed database system having a center system with a link file and

The present invention is a system that converts a vertical file (28) of time series data records from multiple terminals (10-14) and stored by time into a horizontal file (52) where records are organized by terminal (10-14). The user can then access the desired terminal information rapidly in the horizontal file (52) rather than having to perform an inefficient sequential search of the vertical file (28). After the terminal data is horizontally stored, the data is used to create display or print-out windows of the terminal data for many locations. Each window includes the data from a single terminal (10-14). For example, each window could include a graph of rain data for a location. The system also creates a cross reference table (76) of the addresses of the data used to create each window. When the display arrangement is altered, the data in the horizontal file (52/74) is reorganized to match the display order while only the physical addresses of the data for each terminal are changed in a new cross reference table (80).

Sampled data storage and editing system for the time series data being

A data retrieval system for a relational database, for retrieving a relational table in a relational database having parts separately identifiable within the system. A link data file stores link data which include parts descriptors that describe command names assigned to parts located at table item positions of the relational table, attribute descriptors which which describe command names of parts arranged in a row of the relational table corresponding to the command names described by the parts descriptors, and mode descriptors which describe command names of parts arranged in a column of the relational table corresponding to the command names described by the parts descriptors. An actual data file is provided for storing actual data related to the command names of the parts which form the relational table. A link process responds to requests for retrieval from the relational table. The link process obtains a command name corresponding to a retrieval request by referring to the link data file and then obtains the actual data described by the command name, from the actual data file.

System for relational database retrieval through a link file built to

A computing system for a data processing apparatus which carries out data processing using a formula. The system includes a link data file for storing formulas which are described by command names assigned to data elements and operation symbols which define operation between the command names, an actual data file for storing actual data in correspondence with the command names in the formulas managed by the link data file, and a computing file for operation processes for computing the operation contents of the operation symbols in the formulas managed by the link data file. A process is provided for specifying a requested formula which is to be computed by referring to the link data file in response to a computation request, reading from the actual data file the actual data corresponding to the command names in the specified formula and reading from the computing file the operation processes corresponding to the operation symbols in the specified formula, and carrying out operations of the operation processes read from the computing file with respect to the actual data read from the actual data file.

Growing community of inventors

Yokohama, Japan

Tohru Matsumoto