Evaluate Performance in Detail › Performance Evaluation Report › Interpret cluster-wide Summary Statistics

Interpret cluster-wide Summary Statistics

In the Performance Evaluation Report, the summary statistics follow the cluster-wide hot file statistics. The summary statistics highlight the CPU and memory utilization for the configuration.

To display only the summary statistics section of the report, use the following qualifier:

/INCLUDE=SUMMARY_STATISTICS

To disable the summary statistics display from the report, use this qualifier:

/INCLUDE=NOSUMMARY_STATISTICS  

The following example shows the summary statistics section of the Performance Evaluation report:

Performance                          CLUSTER                            Page   1
Evaluation                                                          PA Vx.x
				Tuesday 26-JAN-2006 09:00 to 10:00


  +-----------------------------------------------------------------------+
  | The following table gives a summary of the average CPU and MEMORY     |
  | utilization, and average number of jobs by type for each node.        |
  +-----------------------------------------------------------------------+

Average number of processes
                                  CPU     MEM    ----------------------------
Node        Hardware Type        %Util   %Util   Intractv Batch Netwrk Detach
------   ----------------------   -----   -----   -------- ----- ------ ------
  1          2                      3       4         5       6      7     8
LATOUR   VAX 8700                  29.4    50.3      35.38  0.16   4.13  40.02
 YQUEM   VAX 6000-440              32.4    62.6     115.05  3.00   6.34  41.00
 GALLO   VAX 8700                  27.9    85.1      30.66  1.00   4.48  56.01


User Command: ADVISE PERF REPORT PERF/BEG=26-JAN-2006 09:00:00.00/END=26-JAN-2006 -
10:00:00.00/INCLUDE=SUMMARY/OUT=EXAMPLE_3_13.INFO 
                            9

The following statements are keyed to the previous example:

1. Node name.
2. Type of processor. LATOUR is a VAX 8700, YQUEM is a VAX 6000-400, and GALLO is a VAX 8700.
3. Average percentage of time that each node's CPU was used during the reporting time period. On an SMP system, all active processors are considered when computing this value.
4. Average percentage of each node's memory that was used during the reporting time period.
5. Average number of interactive jobs, by node, during the reporting period.
6. Average number of batch jobs, by node, during the reporting period.
7. Average number of network jobs, by node, during the reporting period.
8. Average number of detached jobs, by node, during the reporting period.
9. Command line used to generate the requested report.