[THIN] KB: CTX114844 - Effects of Varying the Number of CPUs of a Citrix Presentation Server
- From: "Jim Kenzig ThinHelp.com" <jkenzig@xxxxxxxxx>
- To: THIN <thin@xxxxxxxxxxxxx>
- Date: Sun, 7 Oct 2007 11:48:51 -0400
CTX114844 - Effects of Varying the Number of CPUs of a Citrix Presentation
Server
This document was published at: http://support.citrix.com/article/CTX114844
Document ID: *CTX114844*, Created on: Oct 4, 2007, Updated: Oct 5, 2007
Products: Citrix Presentation Server 4.5 for Windows Server 2003 x64
Edition, Citrix Presentation Server 4.5 for Windows Server 2003
*Summary*
The number of users that a Presentation Server can support depends on
several factors including:
- The server's hardware specifications
- The applications used (because of the applications' CPU and memory
requirements)
- The amount of user input being processed by the applications
- What is considered to be maximum desired resource usage on the
server; for example, 90% CPU usage or 80% memory usage?
This article discusses the increase in user capacity when CPUs with multiple
cores are added. First, the Citrix benchmarking test for user capacity,
known as ICAMark, is described.
*Citrix ICAMark*
Citrix ICAMark is an internal tool which is based on the Citrix Server Test
Kit (CSTK) and used by Citrix Engineering for benchmarking purposes to
quantify the optimal number of simulated client sessions that can be
connected to a Presentation Server with acceptable performance. Extending
the number of concurrent simulated users beyond the optimal results will
cause a decrease in performance and may impact end user experience.
The test simulates users constantly typing and performing actions in
Microsoft Excel, Microsoft Access, and Microsoft PowerPoint. Other
applications can utilize more or less memory and CPU than Microsoft Office
and therefore could produce different results. Note also that the simulated
users in this test are constantly typing into these applications and may be
considered more "rigorous" than normal users.
In this test, a step size, "number of users," is defined as 10. During the
course of the test, after the first 10 users are logged in, ICAMark launches
simulated user scripts on all 10 sessions. Each script opens Microsoft
PowerPoint and simulates the creation of a presentation, including copies
and deletion of slides, font changes and presentation viewing. Once the
PowerPoint phase is complete, PowerPoint is closed and Microsoft Access is
opened. The script then simulates the creation of an Access database,
including a table, query, and form, with data manipulation. Once the Access
phase is complete, a Microsoft Excel spreadsheet is created, data populated
into the spreadsheet is used to do number of calculations and charts.
Based on how long the scripts take to complete, an ICAMark score is
calculated. For this test, a score of 80 has been determined as the optimal
load for a server. This means that the server has enough additional CPU and
memory resources to handle spikes in performance. When the test iteration
score drops below 80, additional users added to the server consume more
resources, producing lower test scores and slower performance.
*Number of CPUs Effect on User Capacity*
The benchmark test was performed with the following:
*Server:*
Quad Quad-Core Processors - 3.0GHz Xeon with 8 MB L2 Cache
6x 73 GB 2.5" SAS 10K RPM
64 GB RAM
16 GB Page File
Citrix Presentation Server 4.5
Microsoft Windows Server 2003 64-bit
Microsoft Office 2003 Professional
*Clients:*
Intel XEON 2.8 GHz Processor
1 GB RAM
Citrix Program Neighborhood Version 10.0
Microsoft Windows Server 2003 32-bit
Tests were performed by keeping the hardware static and disabling processors
on the server.
The following results were collected:
*# of CPUs*
*# of Cores*
*# of Simulated Users*
*% Performance Increase*
1
2
202
N. A.
2
4
323
60%
4
8
417
29%
*Number of CPU Cores Effect on User Capacity*
The results conclude that the performance of 8 CPU Cores enabled and 417
concurrent simulated users, is equivalent to the performance of 4 CPU Cores
enabled with 323 concurrent simulated users, which is equivalent to the
performance of 2 CPU Cores enabled with 202 concurrent simulated users.
Moving from a dual to a quad core system equates to a 60% increase in
performance while moving from a quad to an eight core system equates to only
a 29% increase in performance. In other words, as CPU cores are added to the
server, the increase in performance of the operating system becomes less. As
in this scenario server scalability is not linear with the number of
processors, and drops off sharply between 4-8 CPU cores.
All tests were run on Windows Server 2003 64-bit. In this scenario a 32-bit
operating system would be limited by the amount of kernel memory available.
Similar testing was conducted using this scenario on a 32-bit operating
system. The results showed that the system was unable to scale beyond 200
users. This architecture limitation with 32-bit was worked around by using
32-bit and 64-bit applications on a 64-bit operating system.
*Note: *When sizing Presentation Servers the number of actual users per
server varies based on the applications deployed.
*More Information*
See Advanced Concepts Guide - Citrix Presentation Server, Platinum
Edition<http://support.citrix.com/article/entry.jspa?entryID=14748>-
for a list of additional Advanced Concepts Guide articles.
--
Jim Kenzig
Microsoft MVP - Terminal Services
http://www.thinhelp.com
Citrix Technology Professional
Provision Networks VIP
CEO The Kenzig Group
http://www.kenzig.com
Blog: http://www.techblink.com
Other related posts:
