Method for analyzing capacity of parallel processing systems

• US 6,266,804 B1
• Filed: 12/23/1997
• Issued: 07/24/2001
• Est. Priority Date: 12/23/1997
• Status: Expired due to Term

• Current Assignee: Ab Initio Technology LLC (Ab Initio Software Corporation)
• Original Assignee: Ab Initio Software Corporation
• Inventors: Isman, Marshall A.
• Primary Examiner(s): Powell, Mark
• Assistant Examiner(s): ZHEN, WEI Y
• Application Number:
  US08/997,142
• Time in Patent Office: 1,309 Days
• Field of Search: 395/704, 717/4, 717/9, 703/22
• US Class Current: 717/125
• CPC Class Codes: G06F 11/323 Visualisation of programs or trace data
• CPC Class Codes: G06F 11/3404 for parallel or distributed programming
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 2201/88 Monitoring involving counting

Abstract

A method for analyzing the performance and capacity of an application and a parallel processing system. Based upon a graph representation of the application and system and upon supplied performance values, a data file is created which describes the graph and values. Using that data file, performance equations are generated which model the performance of the application and the system. The model is displayed and modified by the user to allow analysis, evaluation, and extrapolation.

Claims

1. A method for analyzing the capacity of an application executing on a parallel processing system and expressed as a graph of vertices, comprising the steps of:

(a) accessing an application for processing designated data records on a parallel processing system, such application being expressed as a graph of vertices;

(b) creating a description of the sizes of data records throughout the graph;

(c) creating a performance description of each vertex in the graph;

(d) determining an execution time for each vertex in the graph;

(e) determining counts of data records assigned to corresponding vertices in the graph; and

(f) creating outputting a description of the total execution time and performance of the parallel processing system based on the determined execution time and counts of data records.

2. The method of claim 1 further comprising the steps of:

(a) creating multiple descriptions of the total execution time and performance of the parallel processing system based on multiple input data sets;

(b) creating a comparison of the multiple descriptions; and

(c) outputting such comparison.

3. A method for analyzing the capacity of an application executing on a parallel processing system and expressed as a graph of vertices and links given a set of supplied values, comprising the steps of:

(a) creating a description of the vertices and links of the graph including connections between vertices and links, data processing rates, and amounts of data;

(b) generating performance characteristics of the application based upon the description, and the set of supplied values, including total execution time, resource requirements, and capacity of the application;

(c) providing a means such that the supplied values can be altered, creating altered values;

(d) re-generating performance characteristics of the application based on the altered values; and

(e) outputting such performance characteristics.

4. The method of claim 3 further comprising the steps of:

(a) accepting multiple sets of supplied values;

(b) generating performance characteristics of the application for each set of supplied values;

(c) calculating sets of estimated values by applying trend equations to the multiple sets of supplied values;

(d) generating performance characteristics of the application based on the estimated values; and

(e) outputting the performance characteristics based on each set of supplied values and based on the estimated values.

5. A method for analyzing the capacity of an application executing on a parallel processing system and expressed as a graph of vertices and links given a set of supplied values, comprising the steps of:

(a) creating a description of the vertices and links of the graph including connections between vertices and links, data processing rates, and amounts of data;

(b) generating performance equations based upon the description which will calculate performance characteristics of the system including total execution time, resource requirements, and capacity of the application;

(c) applying the performance equations to the supplied values;

(d) providing a means such that the supplied values can be altered, creating altered values;

(e) applying the performance equations to the altered values; and

(f) outputting the results of the applied performance equations.

6. The method of claim 5 further comprising the steps of:

(a) accepting multiple sets of supplied values;

(b) applying the performance equations to each set of supplied values;

(c) generating trend equations based upon the multiple sets of supplied values;

(d) calculating sets of estimated values by applying the trend equations to the multiple sets of supplied values;

(e) applying the performance equations to the estimated values; and

(f) providing a means of outputting the supplied values, the estimated values, and stored results.

7. A computer program for analyzing the capacity of an application executing on a parallel processing system and expressed as a graph of vertices and links given a set of supplied values, the computer program being stored on a media readable by a computer system, for configuring the computer system upon being read and executed by the computer system to perform the functions of:

(a) creating a description of the vertices and links of the graph including connections between vertices and links, data processing rates, and amounts of data;

(b) generating performance characteristics of the application based upon the description, and the set of supplied values, including total execution time, resource requirements, and capacity of the application;

(c) providing a means such that the supplied values can be altered, creating altered values;

(d) re-generating performance characteristics of the application based on the altered values; and

(e) outputting such performance characteristics.

8. The computer program of claim 7 further comprising the functions of:

(a) accepting multiple sets of supplied values;

(b) generating performance characteristics of the application for each set of supplied values;

(c) calculating sets of estimated values by applying trend equations to the multiple sets of supplied values;

(d) generating performance characteristics of the application based on the estimated values; and

(e) outputting the performance characteristics based on each set of supplied values and based on the estimated values.

9. A computer-readable storage medium, configured with a computer program for analyzing the capacity of an application executing on a parallel processing system and expressed as a graph of vertices and links given a set of supplied values, where the storage medium so configured causes a computer to operate in a specific and predefined manner to perform the functions of:

(a) creating a description of the vertices and links of the graph including connections between vertices and links, data processing rates, and amounts of data;

(b) generating performance characteristics of the application based upon the description, and the set of supplied values, including total execution time, resource requirements, and capacity of the application;

(c) providing a means such that the supplied values can be altered, creating altered values;

(d) re-generating performance characteristics of the application based on the altered values; and

(e) outputting such performance characteristics.

10. The computer-readable storage medium of claim 9 further comprising the functions of:

(a) accepting multiple sets of supplied values;

(b) generating performance characteristics of the application for each set of supplied values;

(c) calculating sets of estimated values by applying trend equations to the multiple sets of supplied values;

(d) generating performance characteristics of the application based on the estimated values; and

(e) outputting the performance characteristics based on each set of supplied values and based on the estimated values.

11. A computer program, stored on a computer-readable medium, for analyzing the capacity of an application executing on a parallel processing system and expressed as a graph of vertices and links given a set of supplied values, the computer program comprising instructions for causing a computer system to:

(a) create a description of the vertices and links of the graph including connections between vertices and links, data processing rates, and amounts of data;

(b) generate performance characteristics of the application based upon the description, and the set of supplied values, including total execution time, resource requirements, and capacity of the application;

(c) provide a means such that the supplied values can be altered, creating altered values;

(d) re-generate performance characteristics of the application based on the altered values; and

(e) output such performance characteristics.

12. The computer program of claim 11, further comprising instructions for causing the computer to:

(a) accept multiple sets of supplied values;

(b) generate performance characteristics of the application for each set of supplied values;

(c) calculate sets of estimated values by applying trend equations to the multiple sets of supplied values;

(d) generate performance characteristics of the application based on the estimated values; and

(e) output the performance characteristics based on each set of supplied values and based on the estimated values.