Optimizing Windows NT for Performance

 Last reviewed: February 21, 1996
Article ID: Q146005
The information in this article applies to:
  • Microsoft Windows NT Workstation versions 3.5 and 3.51
  • Microsoft Windows NT Server versions 3.5 and 3.51

SUMMARY

When starting troubleshooting performance problems or when trying to optimize Windows NT, while working with Windows NT Performance Monitor it isn't always easy to choose from the large number of performance monitor objects and counters.

This article helps you pick the right counters and objects.

Below you'll find performance checkpoints for the following possible bottlenecks:

  • Suspected bottleneck: memory
  • Suspected bottleneck: processor
  • Suspected bottleneck: physical disk
  • Suspected bottleneck: network
  • Suspected bottleneck: network components

MORE INFORMATION

Suspected Bottleneck: Memory

Check: 

Object:   Memory
Counter: Pages /sec 

Object:   Logical Disk (location of the PAGEFILE.SYS)
Counter: Avg. Disk sec/Transfer

If the product of these two counters (equals percentage of the disk access time used by paging) > 10 % on a sustained basis, the system needs more memory.

Check: 

Object:   Memory
Counter: Pages /sec

If counter value is consistently > 5 , suspect memory.

Check: 

Object:   Server
Counter: Pool Nonpaged Failures

The number of times allocations from nonpaged pool have failed. Indicates that the computer's physical memory is too small.

Check: 

Object:   Server
Counter: Pool Paged Failures

Pool Paged Failures indicate that either physical memory or a paging file is near capacity.

Check: 

Object:   Server
Counter: Pool Nonpaged Peak

The maximum number of bytes of nonpaged pool the server has had in use at any one point. Indicates how much physical memory the computer should have.

Suspected Bottleneck: Processor

Check: 

Object:   Processor
Counter: %Processor Time

If this value is consistently high (> 80%) and disk and network counter values are low, suspect the processor. 

Object:   System
Counter: %Processor Time (for multi processor systems)

If this value is consistently high (> 80%) and disk and network counter values are low, suspect the processor. 

Object:   System
Counter: Processor Queue Length

A sustained processor queue length > 2 , generally indicates a processor bottleneck. 

Object:   Processor
Counter: Interrupts/sec

A dramatic increase in this counter value without a corresponding increase in system activity indicates a hardware problem. 

Object:   Processes (_Total)
Counter: %Processor Time

If more than a couple of processes are contending for the majority of the processor time, then a faster processor or an additional processor should be considered.

Suspected Bottleneck: Physical Disk

Check: 

Object:   Physical Disk
Counter: %Disk Time

If this value is consistently high and disk queue length is greater than 2, suspect the disk. 

Object:   Physical Disk
Counter: Average Disk sec/Transfer

A high value (values greater than 0.3 seconds) may mean that the disk controller is continually retrying the disk because of failures. 

Object:   Physical Disk
Counter: Disk Queue Length Counter: Average Disk sec/Transfer

The Average Queue Time is the average amount of time for a disk transfer (either reads or writes) to complete. Use the following formula to find the average disk queue time: 

   Avg. Queue Time = Disk Queue Length x Avg. Disk sec/Transfer
This information is a relative performance measurement and should be compared with other hard disk drivers in your system. Compute the figures for all logical disks in your system. The number of disk commands waiting in the queue is normally the factor that slows disk performance by increasing the average disk queue time. 

Object:   Physical Disk
Counter: Disk Bytes/sec

A Disk Bytes/sec count lower than 20K may indicate that an application is accessing a disk inefficiently.

Suspected Bottleneck: Network Components

Check: 

Object:   Redirector
Counter: Current commands

If this number is greater than one per network adapter, the redirector may be a bottleneck in the system for one or more of the following reasons:

- the server with which the redirector is communicating is slower than the 

  redirector.
- the network may be experiencing capacity problems. - the redirector is busier than the adapter can keep up with.

If network capacity problems are identified, it may be necessary to subnet the network in an attempt to partition network traffic.

Check: 

Object:   Redirector
Counter: Network Errors/sec

If any network errors are logged, check the Event Log for more details.

Check: 

Object:   Redirector
Counter: Reads Denieds/sec Counter: Writes Denieds/sec

These values indicate if the remote servers are having problems with memory allocation.

Check: 

Object:   Server
Counter: Work Item Shortage

An increase in Work Item Shortage should cause a change in the registry value(s) InitialWorkItems and/or MaximumWorkItems (depending on when the outage occurred).

Check: 

Object:   Server
Counter: Raw Reads Rejected/sec Counter: Raw Writes Rejected/sec

Rejections indicates the exhaustion of RAW work items used when busy doing large file transfers. The increase of the registry value RawWorkItems can possibly solve this bottleneck.

Suspected Bottleneck: Network

Check: 

Object:   Server
Counter: Bytes Total/sec

If the sum of Bytes Total/sec for all servers is roughly equal to the maximum transfer rates of your network, you may need to segment the network.

KBCategory: kbnetwork
KBSubcategory: nthowto ntconfig
Additional reference words: prodnt 3.50 3.51

THE INFORMATION PROVIDED IN THE MICROSOFT KNOWLEDGE BASE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. MICROSOFT DISCLAIMS ALL WARRANTIES, EITHER EXPRESS OR IMPLIED, INCLUDING THE WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL MICROSOFT CORPORATION OR ITS SUPPLIERS BE LIABLE FOR ANY DAMAGES WHATSOEVER INCLUDING DIRECT, INDIRECT, INCIDENTAL, CONSEQUENTIAL, LOSS OF BUSINESS PROFITS OR SPECIAL DAMAGES, EVEN IF MICROSOFT CORPORATION OR ITS SUPPLIERS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR CONSEQUENTIAL OR INCIDENTAL DAMAGES SO THE FOREGOING LIMITATION MAY NOT APPLY.

Last reviewed: February 21, 1996
© 1998 Microsoft Corporation. All rights reserved. Terms of Use.