HP 3000 Manuals

Memory Detail Screen [ HP GlancePlus User's Manual (for MPE/iX Systems) ] MPE/iX 5.0 Documentation


HP GlancePlus User's Manual (for MPE/iX Systems)

Memory Detail Screen 

The Memory Detail screen displays metrics that show physical memory
activity by the system during the current interval.  The detailed data
area of the display is divided into two parts:

   *   Fault Types.

   *   Miscellaneous.

[]
Figure 5-23. Memory Detail Screen Fault Types The Fault Types section presents several metrics for memory faults (faults per second). [Scale: 0 to 50 faults per second.]
[]
Figure 5-24. Memory Detail Screen--Fault Types The Fault Types are as follows: Transient Unspecified transient data objects. Permanent Unspecified permanent data objects. NM Stack Native Mode process stacks. CM Stack Compatibility Mode process stacks. NM Code Native Mode code. CM Code Compatibility Mode code (program files). CM Data Compatibility Mode data segment. File Object File Objects. NM System Native Mode system library. Library CM System Compatibility Mode system library. Library Miscellaneous The Miscellaneous section presents several system memory performance metrics that could be used to indicate potential memory bottlenecks. A bottleneck occurs when a memory resource has reached its capacity limit and might slow system throughput or response time.
[]
Figure 5-25. Memory Detail Screen--Miscellaneous Metrics MemMgr Memory Manager utilization during the current time interval is shown on the bar and by percentage to the right of this field. Sustained utilization of 8% or greater may indicate a potential memory bottleneck. Page Overlay This metric, depicted as the number of Candidate Rate potential free pages per second, shows the Memory Manager activity on the making of physical memory pages as Overlay candidates (potential free pages). A sustained rate greater than 35 free pages per second may indicate a memory bottleneck. Table 5-2. (cont.) Memory Mgr The bottom line of the Memory Detail screen Clock Cycle Rate shows the memory manager clock cycle rate. The clock cycle rate is computed by dividing the number of times the memory manager has cycled through memory by the elapsed time GlancePlus has been running. Unlike the other metrics on the screen, this one is relative to the run time of GlancePlus rather than the current interval. This avoids the misrepresentation of reporting one clock cycle over four thirty-second intervals as a rate of 120 per second during one interval and 0 per second during the other three. It also means that you should be careful to let GlancePlus run a few minutes before drawing any conclusions from this metric. A memory manager clock cycle rate of more than 25 per hour over ten minutes or more may indicate possible memory pressure. Page Fault Rate This field shows the sum of the Global System Page Fault types. The sum is expressed in page faults per second. A fault rate greater than 30 faults per second sustained over several intervals may indicate a potential bottleneck. System Library This metric shows the total Native Mode and Page Faults Compatibility Mode system library faults. A sustained rate of greater than 2 faults per second may indicate a memory bottleneck condition. Physical Memory This metric indicates the size of the Size machine's physical memory (in megabytes).


MPE/iX 5.0 Documentation