Real time internal resource monitor for data processing system

• US 6,049,798 A
• Filed: 06/14/1994
• Issued: 04/11/2000
• Est. Priority Date: 06/10/1991
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Shay, Robert Charles, Yeamans, Stephen Asa, Holck, Timothy Manfred, Raby, Charles Lincoln, Knox, Telford Jr., Bishop, David Addy, Turner, Mark David
• Primary Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US08/259,368
• Time in Patent Office: 2,128 Days
• Field of Search: 355/650, 355/200, 355/700, 355/500, 707/10
• US Class Current: 1/1
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3423 where the assessed time is active or idle time
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 11/3485 for I/O devices
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 2201/805 Real-time
• CPC Class Codes: G06F 2201/835 Timestamp
• CPC Class Codes: G06F 2201/86 Event-based monitoring
• CPC Class Codes: G06F 2201/88 Monitoring involving counting

Abstract

A system resource monitor is provided to capture a data processing system'"'"'s internal resource utilization, such as memory, CPU, or peripheral device availability/utilization. The captured `state` of the data processing system'"'"'s resources is maintained in real-time, while the impact on the system'"'"'s performance in providing such information is kept to a minimum. This is accomplished through a combination of various techniques, including specialized device drivers for the respective devices coupled with a unique data reduction technique. Such techniques include filtering only events which are of interest and combining similarly related events to reduce data processing requirements. This real-time support provides an immediate and accurate representation of the internal operations of the data processing system. Further, these resources can monitored at the process level of a multiprocessing system. This captured data can be used by an application or control program to identify, isolate, and fine-tune the data processing system'"'"'s resources to improve the overall efficiency of the system being monitored.

Claims

1. A system for monitoring the performance of one or more computer processors and a plurality of processor resources assigned to said one or more processors, said one or more computer processors and including permanent storage devices and volatile storage devices, the system comprising:

control program means executing on each of said one or more computer processors for generating event signals at a start and an end of predefined events;

monitor control means for controlling collection of performance data and for displaying performance data;

data collection means for collecting performance data from said one or more computer processors, the data collection means being responsive to said monitor control means, said data collection means further comprising;

storage means for storing performance data;

processor resource monitor means for testing the status of said processor resources comprising;

Polling means for polling permanent and volatile storage status; and

data interpretation means for generating performance data based on said polling means;

event signal filter means for testing each event signal, and writing to said storage means only that data associated with a selected subset of event signals.

2. A method for tracing system events in a processor system having one or more processing elements and a hardware timer, the method comprising the steps of:

executing at least one program in said processor system that generates system event signals;

initiating system event monitoring;

capturing selected ones of said system event signals, said selected ones defined by a control processor;

generating a time stamp from said hardware timer indicating a time of occurrence;

matching a start event signal with a corresponding end event signal to create an event interval signal;

determining event interval time by analyzing said time stamps for said start and end event signals;

summarizing said event interval signals to create system event performance data; and

displaying system event performance data in a graphical form.

3. A method for determining amount of volatile memory in use over a time interval in a computer system having at least one processor and volatile memory divided into a plurality of allocatable memory blocks each having a block size, each of said memory blocks containing an indicator of whether or not said block is allocated and a time stamp indicating the time of last processor access to said block, said method comprising the steps of:

periodically testing each of said memory blocks to determine whether said memory block is allocated or not;

if allocated, testing said time stamp to determine the time of last access;

testing to determine whether said last access is within said monitor time interval; and

adding the block size to a counter summing the total memory in use during said time interval.

4. A method of measuring the performance of a computer system having one or more processors, a hardware timer, and a plurality of processor resources including memory divided into a plurality of blocks each having a block size, the method comprising the steps of:

executing at least one program on said computer system, said at least one program generating system event signals;

summarizing system event performance by determining event interval times for a selected subset of system events;

summarizing processor idle time by determining the percentage of time each of said one or more processors is not executing a task of higher than a lowest available priority;

summarizing processor resource utilization by determining the percentage of time the resource is busy in a time interval;

summarizing memory usage by determining the amount of memory used over a time interval, and

collecting all said summaries and displaying said data on a display device.

5. The method of claim 4, wherein the process for summarizing system event performance comprises the steps of:

receiving said event signals;

testing each event signal and saving only selected event signals contained in a predefined subset of event signals;

storing said selected event signals; and

processing said selected events signals to generate an event summary.

6. The method of claim 4, wherein the process for summarizing system event performance comprises the steps of:

initiating system event monitoring;

capturing selected ones of said system event signals, said selected ones defined by a control processor;

generating a time stamp from said hardware timer indicating a time of occurrence;

matching a start event signal with a corresponding end event signal to create an event interval signal;

determining event interval time by analyzing said high resolution time stamps for said start and end event signals; and

summarizing said event interval signals to create system event performance data.

7. The method of claim 4, wherein the process of summarizing processor idle time comprises the steps of:

starting a first monitoring process;

assigning said first monitoring process said lowest available priority;

monitoring by a second monitoring process an amount of time spent executing said monitoring process; and

reporting said amount of time spent as processor idle time.

8. The method of claim 4, wherein the process of summarizing memory usage comprises the steps of:

periodically testing each of said memory blocks to determine whether said memory block is allocated or not;

if allocated, testing said time stamp to determine the time of last access;

testing to determine whether said last access is within said monitor time interval; and

adding the block size to a counter summing the total memory in use during said time interval.