LabVIEW Real-Time

Course Manual

November 2001 Edition

Part Number 323226A-01

© 2001 National Instruments Corporation.

All Rights Reserved.
 Copyright
Under the copyright laws, this publication may not be reproduced or transmitted in any form, electronic or
mechanical, including photocopying, recording, storing in an information retrieval system, or translating, in whole
or in part, without the prior written consent of National Instruments Corporation.

Trademarks
DataSocketTM, FieldPointTM, LabVIEWTM, National InstrumentsTM, NI-CANTM, ni.comTM, NI-DAQTM, NI-MotionTM,
and NI-VISATM are trademarks of National Instruments Corporation.

Product and company names mentioned herein are trademarks or trade names of their respective companies.

Worldwide Technical Support and Product Information

ni.com

National Instruments Corporate Headquarters

11500 North Mopac Expressway Austin, Texas 78759-3504 USA Tel: 512 683 0100

Worldwide Offices
Australia 03 9879 5166, Austria 0662 45 79 90 0, Belgium 02 757 00 20, Brazil 011 284 5011,
Canada (Calgary) 403 274 9391, Canada (Montreal) 514 288 5722, Canada (Ottawa) 613 233 5949,
Canada (Québec) 514 694 8521, Canada (Toronto) 905 785 0085, China (Shanghai) 021 6555 7838,
China (ShenZhen) 0755 3904939, Czech Republic 02 2423 5774, Denmark 45 76 26 00, Finland 09 725 725 11,
France 01 48 14 24 24, Germany 089 741 31 30, Greece 30 1 42 96 427, Hong Kong 2645 3186,
India 91805275406, Israel 03 6120092, Italy 02 413091, Japan 03 5472 2970, Korea 02 596 7456,
Malaysia 603 9596711, Mexico 001 800 010 0793, Netherlands 0348 433466, New Zealand 09 914 0488,
Norway 32 27 73 00, Poland 0 22 528 94 06, Portugal 351 1 726 9011, Russia 095 2387139, Singapore 2265886,
Slovenia 386 3 425 4200, South Africa 11 805 8197, Spain 91 640 0085, Sweden 08 587 895 00,
Switzerland 056 200 51 51, Taiwan 02 2528 7227, United Kingdom 01635 523545
Table of Contents
Preface..............................................................................................................................................1
Lesson 1 LabVIEW Real-Time Basics ........................................................................................ 1-1
Exercise 1-1 ............................................................................................................................... 1-23
Exercise 1-2 ............................................................................................................................... 1-23
Exercise 1-3 ............................................................................................................................... 1-24
Lesson 2 Communication with RT Series Hardware................................................................... 2-1
Exercise 2-1 ............................................................................................................................... 2-42
Exercise 2-2 ............................................................................................................................... 2-46
Exercise 2-3 ............................................................................................................................... 2-47
Exercise 2-4 ............................................................................................................................... 2-48
Exercise 2-5 ............................................................................................................................... 2-49
Exercise 2-6 ............................................................................................................................... 2-49
Lesson 3 Multithreading and Priorities........................................................................................ 3-1
Exercise 3-1 ............................................................................................................................... 3-23
Exercise 3-2 ............................................................................................................................... 3-23
Exercise 3-3 ............................................................................................................................... 3-26
Exercise 3-4 ............................................................................................................................... 3-28
Lesson 4 Timing and Sleep.......................................................................................................... 4-1
Exercise 4-1 ............................................................................................................................... 4-27
Exercise 4-2 ............................................................................................................................... 4-28
Exercise 4-3 ............................................................................................................................... 4-29
Lesson 5 Shared Resources.......................................................................................................... 5-1
Exercise 5-1 ............................................................................................................................... 5-24
Exercise 5-2 ............................................................................................................................... 5-27
Exercise 5-3 ............................................................................................................................... 5-30
Lesson 6 Other Features of LabVIEW Real-Time ...................................................................... 6-1
Exercise 6-1 ............................................................................................................................... 6-24
Exercise 6-2 ............................................................................................................................... 6-28
Exercise 6-3 ............................................................................................................................... 6-29
Exercise 6-4 ............................................................................................................................... 6-30
Exercise 6-5 ............................................................................................................................... 6-32

iii
 LabVIEW Real-Time

ni.com

© National Instruments Corporation 1 LabVIEW Real-Time

 Course Goals

! Understand concepts of real-time and determinism

! Learn how to configure and communicate with
real-time hardware
! Understand memory usage in LabVIEW Real-Time
! Understand how multithreading and priorities work
in LabVIEW Real-Time
! Understand shared resources
! Learn how to optimize real-time tasks
! Learn about other real-time features

ni.com

As a prerequisite to this course, you should have at least taken the LabVIEW
Basics I course.

LabVIEW Real-Time 2 © National Instruments Corporation

 Course Non-Goals

! To teach LabVIEW basics

! To teach control theory
! To teach PID and Fuzzy Logic theory
! Analog-to-digital (A/D) theory
! The operation of GPIB, RS-232, Motion, CAN, VISA
! To discuss every built-in LabVIEW object, function,
or library VI
! The development of a complete application for any
student in the class

ni.com

© National Instruments Corporation 3 LabVIEW Real-Time

 Equipment Needed for This Course

!Host PC running Windows 95 or newer

!LabVIEW Real-Time version 6i or newer
!LabVIEW Real-Time Course Manual
and disks
!PXI-MIO and PCI/PXI-7030 MIO devices
!PXI-8156B, 8170, 8145, or 8146 controller
!DAQ signal accessory
!FP-2000 module
ni.com

To complete the exercises in this course, one of the following sets of hardware is required:
• PXI controller (8156B, 8170 with ethernet module 8211, 8145, 8146)
• PXI-MIO board
• PXI chassis
• ethernet or crossover cable
• host PC
OR
• PCI/PXI RT Series DAQ board
OR
• host PC
• ethernet or crossover cable
• FP-2000
Some exercises are hardware specific, so you may not be able to complete all the exercises
with your hardware set.

LabVIEW Real-Time 4 © National Instruments Corporation

 Hands-On Exercises

! Exercises reinforce
the topics presented
! Each exercise shows
a finished VI after it
is run
! Descriptions of how to
build the VI follow the
picture
! Save exercises
in personal folder on
the Windows desktop

ni.com

© National Instruments Corporation 5 LabVIEW Real-Time

 Course Outline

! Real-time concepts and definitions

! Real-time OS and real-time targets
! Communicating with the real-time targets
! Multithreading and priorities
! Shared resources
! Hardware timing
! Control loop optimization
! Watchdogs
! Application Builder

ni.com

LabVIEW Real-Time 6 © National Instruments Corporation

 Course Outline
DAY 1 DAY 2

Real-Time Concepts Shared Resources

Real-Time Definitions Hardware Timing
Hardware Targets Loop Optimization
Communication Watchdogs
Multithreading Application Builder
Priorities

ni.com

This course will focus on LabVIEW Real-Time and DAQ. Although Serial, GPIB,
and VISA are supported in LabVIEW Real-Time 6, they will not be discussed.
To learn how to program with these protocols in LabVIEW, take the necessary
courses.

© National Instruments Corporation 7 LabVIEW Real-Time

 Lesson 1

LabVIEW Real-Time Basics

You Will Learn:

A. Real-time definitions and concepts
B. About the real-time targets
C. About the real-time OS
D. About differences between LabVIEW
and LabVIEW Real-Time

ni.com

© National Instruments Corporation 1-1 LabVIEW Real-Time

 What Is Real-Time?

Real-time response is the ability to reliably,

without fail, respond to an event, or perform
an operation, within a guaranteed time
period.

ni.com

Before delving into how to use LabVIEW Real-Time, it is first important to understand
some technical terms related to real-time systems. The first term to understand is what is
meant by real-time. A misconception is that it means really fast. Real-time is probably
more synonymous to in-time. In other words, a real-time system ensures that responses
will occur in time, or on time. With non-real-time systems, you do not have any way to
ensure that your response will occur within any given time period, and calculations may
finish much later or earlier than you expect them to.
In order for a system to be a real-time system, all parts of it need to be real-time.
For instance, even though a program runs in a real-time operating system, it does not
mean that the program will behave with real-time characteristics. The program may rely
on something that does not behave in real-time, which then causes the program
to not behave in real-time.

LabVIEW Real-Time 1-2 © National Instruments Corporation

 What Is the Goal of Real-Time?

Deterministic loops with I/O (the focus of this

course will be I/O using NI-DAQ)

Input
Time critical
Control Algorithm
or other Real-Time
code

Output
ni.com

© National Instruments Corporation 1-3 LabVIEW Real-Time

 Real-Time Definitions

!Determinism: A characteristic of a system that

describes how reliably it can respond to events,
or perform operations, within a given time limit.
!Loop Cycle Time: The time taken to execute
one cycle of a loop.
!Jitter: The variation of loop cycle time from the
desired loop cycle time.
• Bounded in real-time systems
• Un-bounded or very large in non-real-time systems

ni.com

The first term to understand is determinism. Determinism is a characteristic

of a system that describes how reliably it can respond to external events, or perform
operations, within a given time limit. High determinism is a characteristic of real-time
systems, which guarantees that your calculations and operations will occur within a given
time. Deterministic systems are predictable. This is an important characteristic for
applications such as control. In a control application, the control program measures inputs,
makes calculations based on the inputs, and then outputs values that are a result of those
calculations. Real-time systems can guarantee that the calculations will finish on time,
all of the time.
Many applications that require a real-time operating system are cyclic, such as a control
application. The time between the start of each cycle, T, is called the loop cycle time
(or sample period). 1/T is the loop rate or sample rate. Even with real-time operating
systems, the loop cycle time can vary between cycles. The amount of time that the loop
cycle time varies from the desired time is called jitter. The maximum amount that a loop
cycle time varies from the desired loop cycle time is called maximum jitter.

LabVIEW Real-Time 1-4 © National Instruments Corporation

 Maximum Jitter
Desired Loop Time
Maximum
Jitter
1
Loop Iteration

Loop Time
ni.com

All real-time systems have some jitter, it is just that the jitter is much smaller than
in non-real-time systems, sometimes on the order of nanoseconds. Non-real-time
systems have very large or unbounded maximum jitter.

© National Instruments Corporation 1-5 LabVIEW Real-Time

 Why Windows Is Not Real-Time

!Processor time shared between programs

!Windows can preempt high priority VIs
!Many programs run in the background
(Screen savers, disk utilities, virus software, etc.)
!Windows must service interrupts
(Keyboard, mouse, Ethernet, etc.)
!Cannot guarantee determinism

ni.com

LabVIEW applications running in Windows are not guaranteed to run in real-time.

This is because Windows is not a real-time operating system. Windows cannot give
you the guarantee that your code will always finish within given time limits. The
time your code takes to execute in Windows is dependant on many factors, such as
other programs running, like screen saver or virus software. Windows also has to
service interrupts (from devices such as a USB port, keyboard, mouse, and other
peripherals), which can delay execution of your code. You can increase your
chances of having programs run deterministically in Windows by disabling all other
programs such as screen savers, disk utilities, and virus software. You can further
increase determinism by disabling drivers for devices with interrupts such as the
keyboard, mouse, and Ethernet card. And finally, for better determinism, you can
write a device driver in Windows to get the most direct access to hardware possible.
Even after all these attempts to increase determinism, there is still no guarantee that
your code will execute with real-time behavior all of the time. This is because
Windows can preempt your LabVIEW programs, even if you use time-critical
priority. Priority will be discussed later.
Fortunately, with LabVIEW Real-Time, your programs run in a separate Real-Time
operating system, so you do not need to disable programs or write device drivers in
order to achieve real-time performance.

LabVIEW Real-Time 1-6 © National Instruments Corporation

 Real-Time DAQ and Windows

Buffered analog I/O can occur in real-time

with Windows (output does not depend on
input)
• DMA
• Hardware clock

ni.com

If all you need to do is real-time data acquisition, then you do not necessarily need
LabVIEW Real-Time. National Instruments has many data acquisition boards that
can acquire data in real-time even though they are controlled by programs running
in Windows. This is done using an onboard hardware clock that makes sure that data
is acquired at constant rates. This data is automatically transferred directly to
memory using DMA or interrupt service requests (IRQ).

© National Instruments Corporation 1-7 LabVIEW Real-Time

 What Is LabVIEW Real-Time (RT)?
An extension of LabVIEW that enables easy
development of real-time applications that run
on National Instruments RT Series Hardware
!LabVIEW Real-Time 5.1.2 was a full
installation that included all of the functionality
of LabVIEW 5.1.1.
!LabVIEW Real-Time 6i is a module that installs
on top of LabVIEW 6i.

ni.com

LabVIEW Real-Time 1-8 © National Instruments Corporation

 Exercise 1.1
Students become familiar with the
documentation available with LabVIEW
Real-Time.

Time to complete: 10 minutes

ni.com

© National Instruments Corporation 1-9 LabVIEW Real-Time

 Real-Time System Overview

!Components required:
• Windows PC (Host PC)
• LabVIEW Real-Time Development System
(on Host PC)
• National Instruments Real-Time Series Hardware
– Separate CPU and memory
– Separate real-time kernel (OS)
• Cables and Accessories
!Same LabVIEW programming techniques

ni.com

All you need to create a real-time system is the following:

• Computer running Windows 2000/NT/9x.
• LabVIEW Real-Time Software
• National Instruments Real-Time Series Hardware
• Cables and accessories
A programmer who already knows how to program using LabVIEW does not have
much more to learn in order to create real-time applications. There is very little
difference between programming in LabVIEW for Windows and LabVIEW Real-
Time. In fact, many existing LabVIEW for Windows programs can be automatically
downloaded and run on National Instruments Real-Time Series Hardware with little
or no modification. However, to get optimal performance, good programming
techniques, as discussed in this course, are required.

LabVIEW Real-Time 1-10 © National Instruments Corporation

 Software Paradigm
Host PC running
Windows and
LabVIEW Real-Time
Development System

Communication

RT Engine
RT Series Hardware (embedded version
Target Platform running of LabVIEW)
RTOS and RT Engine
(Headless) RTOS

ni.com

LabVIEW Real-Time applications are developed in the Windows environment on

a computer called the Host PC, and are then downloaded to National Instruments
Real-Time hardware, where they run in real-time, independent of the Windows OS.
There are several kinds of National Instruments real-time hardware platforms upon
which real-time applications can run. These platforms are also referred to
as targets, or target platforms. The software that runs on the target platforms is
referred to as the LabVIEW RT Engine, or just the RT Engine. The RT Engine
is the software that runs in the real-time operating system (RTOS) on RT Series
Hardware.

© National Instruments Corporation 1-11 LabVIEW Real-Time

 Why LabVIEW Real-Time?

!Programs run in a separate real-time

operating system on RT Series Hardware.
!Real-Time operating system only runs the
RT Engine and no other programs for high
determinism. *
!Less software components results in
reliability and stability.

* Programs related to the RT Engine run at normal priority such

as the FTP Server.

ni.com

There are many applications for which LabVIEW Real-Time is needed instead of
LabVIEW for Windows. These are applications that require programs to run with
a level of determinism unattainable with Windows.
One advantage is that when an RT application is running on RT Series hardware,
if the Host PC crashes, the RT Engine continues to run. Also, since all that is
running on the RT Series Hardware is the LabVIEW RT Engine, it is much more
stable than Windows, which combines software and drivers from many vendors.
Even though your programs are running in a real-time operating system, it does not
mean that they are not susceptible to poor performance due to poor programming
techniques. A well-programmed loop will have higher loop rates and less jitter than
a poorly programmed loop. There are cases where poor programming will cause
your program to crash. For instance, if you make a program that continues to use
more and more memory, as soon as all of the memory is used up, the program will
crash (no virtual memory is used in the RTOS because it is not deterministic). Also,
if you make your program execution dependant upon non-deterministic external
processes, such as communication (i.e., GPIB, Serial, or Ethernet), or File I/O, then
your program is not guaranteed to run in real-time.

LabVIEW Real-Time 1-12 © National Instruments Corporation

 Real-Time Series Hardware

RT Series DAQ Devices FP-2000

(PCI/PXI-7030) Series

PXI Controllers

ni.com

One type of target or platform is a device that plugs into a PCI or PXI slot
of a computer running Windows. It is referred to as an RT Series DAQ Device.
Another type of platform is the RT Series PXI Controller. There is also
a FieldPoint module that can run the RT Engine.

© National Instruments Corporation 1-13 LabVIEW Real-Time

 RT Series DAQ Devices
!RTOS runs on processor board
!DAQ Daughterboard on embedded PCI bus
!No hard drive
DAQ Daughterboard PXI Version
Memory

Processor
Processor Board
ni.com

RT Series DAQ Devices contain a processor board and an attached daughterboard

for I/O. The processor board is a self-contained computer on one board, which runs
the RT Engine. RT Series DAQ Devices communicate with the Host PC using shared
memory located on the RT Series DAQ Device.
Currently, three RT Series DAQ Devices are available:
• PXI/PCI-7030/6030
• PXI/PCI-7030/6040
• PXI/PCI-7030/6533

LabVIEW Real-Time 1-14 © National Instruments Corporation

 RT Series PXI Controller
!Controls devices in chassis in real-time
!Headless
• Keyboard can be used for bios configuration
• Monitor provides limited information

• Hard Drive or CompactFlash

• Processor
• Memory (expandable)
• Serial, Parallel
• GPIB (select controllers)
• Ethernet (separate module
with PXI-8170)

ni.com

The RT Series PXI Controllers come with the RT Engine preinstalled. During
development, RT Series PXI Controllers communicate with the Host PC through
an Ethernet interface. Since RT Series PXI Controllers are controlled remotely,
they are considered headless systems. A headless system has no keyboard, monitor,
or mouse.
As of September 25, 2001, these are the RT Series PXI Controllers available:
• PXI-8156B
• PXI-8170 (requires a separate Network module)
• PXI-8145
• PXI-8146

© National Instruments Corporation 1-15 LabVIEW Real-Time

 FP-2000 Series
Dedicated 486 66MHz
AMD processor running
real-time kernel
Watchdog Timer
User-defined dip
switches and LEDs
On-board static memory
Backup power supply
Serial port
Ethernet port

ni.com

The FP-20XX has many features that enable it to act as a stand-alone device. Since most stand-
alone devices reside in remote or inaccessible locations, it is important for these systems to be
reliable to eliminate the need for constant maintenance. The dedicated processor in the FP-20XX
runs a real-time operating system which executes the LabVIEW Real-Time application
deterministically. The reliability of the real-time operating system also reduces the chances
of system malfunction.
The FP-20XX includes a Watchdog timer that monitors the overall operation of the FieldPoint
system. When the watchdog senses that the system has a malfunction, it can run a separate
maintenance routine or even reboot the FieldPoint system. The operation of the watchdog
is completely user-defined through a simple and intuitive protocol.
The FP-20XX also includes 5 DIP switches, 3 tri-color LEDs, and 1 bi-color LED that can be
used for user interaction as a stand-alone system. For instance, the LEDs can be used to signify
certain states of operation while the dip switches can be configured to specify the startup state
of the FieldPoint system.
Switches 1, 2, 3, 4, and 5 are user-accessible DIP switches. They may be read by the FieldPoint
LabVIEW VIs, but have no default functionality. The DISABLE VI, SAFE MODE, and RESET
DIP switches are read only when you power up or reboot the FP-20XX network module.
You must reboot the module with one of these switches ON for its setting to take effect. If you
reboot the system with more than one of these switches ON, only the left-most switch is read,
and the others are ignored.
The onboard static memory consists of 3 MB of Flash memory for the FP-2000 and 11 MB of
Flash memory for the FP-2010. The Flash memory acts as a hard drive for the system and can
be used to store startup applications or data. For instance, you can configure the FP-2000 to
monitor the temperature inside an environmental chamber while logging that data to the Flash
memory. The FP-2010 is ideal for large data-logging applications with the larger 11 MB of
Flash memory. If you need to stream to disk larger memory chunks, you can transfer data back
to the host program, and stream it to the host PC hard drive.

LabVIEW Real-Time 1-16 © National Instruments Corporation

 FP-2000 Series (Continued)

ni.com

The FP-20XX includes a second set of screw terminals to connect a backup power supply.
The system will automatically switch to this second power supply, when it senses a problem
with the first power supply.
The serial port allows the FP-20XX to connect to existing systems. In addition, it can be used
to connect other serial devices such as a serial LED which can be used to display information
for the operator.
The FieldPoint I/O modules are compact modules ideal for usage in industrial environments,
where space is limited, or where placing a standard desktop computer is not an option.
The dimensions of the module are 4.3 in. wide, 4.3 in. deep, and 3.6 in. high. The systems
can be din-rail mounted to increase space on the factory floor, and be placed in locations with
stringent temperature demands. A FieldPoint 2000 based system can operate efficiently in
environments from –25 to 55 ºC.
With LabVIEW Real-Time downloading the code into the FP-20XX, you can create a stand-alone
system. For instance, you can create a custom data-logging application that acquires the data from
the FieldPoint I/O module and logs the data to the local Flash memory. With over 100 analysis
routines such as digital filters, spectrum analysis, signal processing, and curve fitting in LabVIEW,
you can include custom in-line data analysis on the FP-20XX node. The in-line analysis can
convert the raw data into more effective information by determining and logging the peak
frequency of a signal which also reduces the amount of information stored on the local Flash
memory.
You can also create a stand-alone control application that continually monitors and responds to
a set of stimuli. For instance, this system can be used to monitor and control the pressure in an
environmental chamber. With built-in PID, fuzzy logic, and other control algorithms, you can
build a custom control routine for the system. With this custom control routine running in the
stand-alone control node, the FP-20XX has device level intelligence that doesn’t rely on a main
supervisory control station to make decisions for it. In addition to being more reliable, this
improves the performance of the system with a real-time system responding deterministically
to stimuli.
© National Instruments Corporation 1-17 LabVIEW Real-Time
 LabVIEW RT Menu Items and VIs

!Additional Real-Time menu items

• File»Exit without closing RT Engine VIs
• Operate»Download Application
• Operate»RT Engine Info
• Operate»Switch Execution Target…
• Tools»RT Target Options…
• Help»View Printed Real-Time Manuals…
• Help»LabVIEW Real-Time Help…
!Just a few RT-specific VIs

ni.com

LabVIEW Real-Time 1-18 © National Instruments Corporation

 Configuration with MAX
!MAX is used closely with LabVIEW Real-Time
to configure RT Series Hardware
!New Features for Real-Time

ni.com

New features have been added to Measurement & Automation Explorer (MAX).
Functionality that was previously found in Remote System Explorer (used with
LabVIEW Real-Time 5.1.2) is now found in MAX. In the Configuration window,
you will find an entry called Remote Systems. It is here that you configure RT Series
Devices on the network. The use of MAX will be discussed later.
Fieldpoint Explorer 3.0 is used to configure Fieldpoint 2000 Series modules. See the
user manual for details.

© National Instruments Corporation 1-19 LabVIEW Real-Time

 LabVIEW RT Examples Directory

\Examples\RT\

ni.com

LabVIEW Real-Time 1-20 © National Instruments Corporation

 Exercises 1.2 and 1.3

!Students become familiar with the

documentation available with MAX.
!Students become familiar with the examples
specific to LabVIEW Real-Time.

Time to complete: 10 minutes

ni.com

© National Instruments Corporation 1-21 LabVIEW Real-Time

 Chapter 1—Summary

!LabVIEW Real-Time for Determinism

!Development in Windows similar to LabVIEW
!Code downloaded to RT Engine
!RT Engine runs in RTOS on RT Series
Hardware
• RT Series DAQ Devices
• RT Series PXI Controllers
• FP-2000 Series

ni.com

LabVIEW Real-Time 1-22 © National Instruments Corporation

 Exercise 1-1
Objective: To become familiar with the documentation available with LabVIEW
Real-Time.

1. Launch LabVIEW Real-Time by selecting Start»Programs»National Instruments»

LabVIEW 6»LabVIEW. (If a window appears titled Please select target platform,
select Host PC (LabVIEW for Windows), and click OK.)
2. Click on the New VI button.
3. From the new VI, select Help»View Printed Real-Time Manuals
(be careful that you don’t accidentally select View Printed Manuals…).
4. Take about 5 minutes to browse these documents related to LabVIEW Real-Time:
• LabVIEW Real-Time Release Notes
• LabVIEW Real-Time User Manual
• RT Series DAQ Device User Manual
• RT Series PXI/CompactPCI Controller User Manual
5. From the new VI, select Help»LabVIEW Real-Time Help…
6. Take a few minutes to browse these help files.
7. Exit LabVIEW.

Exercise 1-2
Objective: To become familiar with the documentation available with MAX.

1. Launch MAX by selecting Start»Programs»National Instruments»

Measurement & Automation Explorer.
2. If the User Preferences Dialog Box appears, select Every time I launch MAX,
and uncheck the Show this dialog the next time I launch MAX box.

3. Select Help»Help Topics»Remote Systems.

4. Take a few minutes to browse these help files.
5. Close the help window and MAX.

 National Instruments Corporation 1-23 LabVIEW Real-Time

 Exercise 1-3
Objective: To become familiar with the examples specific to LabVIEW Real-Time.

1. Launch LabVIEW.
2. Click on the Open VI button and browse to the c:\Program Files\National
Instruments\LabVIEW 6\Examples\RT directory.
3. Open each *.llb file and browse the contents of each.

Most examples in these libraries contain two VIs, one that runs on the RT Engine,
and a corresponding VI that runs on the Host PC. The RT Engine VI runs in real-time
and communicates with the Host PC VI, which provides a User Interface.

4. After browsing for a few minutes, exit LabVIEW.

LabVIEW Real-Time 1-24  National Instruments Corporation

 Lesson 2

Communication with RT Series Hardware

You Will Learn:

A. How to configure, target, and download
programs to RT Series Hardware
B. How to communicate with the RT Engine

ni.com

This lesson will briefly describe how to configure RT Series Hardware, but it will
not go into specifics of installation or configuration of either software or hardware.
This information is covered in detail in the user manuals. This lesson will also
discuss communication methods between the Host PC and the RT Engine.

© National Instruments Corporation 2-1 LabVIEW Real-Time

 Configuring 7030s in MAX

• View»Refresh, or <F5>
• Right-click on 7030, select Properties…
– Test Resources
– You can change Device Number
– No Test Panels (grayed out)
ni.com

NI-DAQ software should be installed before installing RT Series DAQ Devices.

If your device does not show up in MAX, selecting View»Refresh, or pressing
<F5> will refresh the Configuration window. If it still does not show up, review
and repeat the installation procedure in the RT Series DAQ Device User Manual.
The Test Resources button will test to make sure that the I/O address, Interrupt,
and DMA channels for your RT Series DAQ Device are properly assigned and
functional.
Because the processor board is not a DAQ board, it has no test panels.

LabVIEW Real-Time 2-2 © National Instruments Corporation

 Configuring Daughterboards

• Right-click on
Daughterboard,
select Properties…
– You can change Device Number
– No Test Resources or Panels (grayed out)
– Select Accessory
ni.com

You cannot test resources or run test panels for the DAQ daughterboard because,
even though the daughterboard is configured using the Host PC, the daughterboard is
actually located on the local PCI bus of the RT Series processor board. Only LabVIEW
Real-Time applications targeted to the RT Series DAQ Device can access the DAQ
daughterboard. Applications run on the Host PC cannot access the DAQ daughterboard
directly, even though the DAQ daughterboard is configured as a separate device.
For this reason, the resource configuration on the PCI bus cannot be tested from
the Host PC. Note that you do not need to test the resource configuration because
National Instruments configures it and it does not change.

© National Instruments Corporation 2-3 LabVIEW Real-Time

 RT Series DAQ Configuration

Create Virtual Channels for RT Series DAQ

Devices from MAX:
• My System»Data Neighborhood»Create
New»Virtual Channel
• Must reset RT Series DAQ Device for virtual
channels to take effect—resetting discussed later

ni.com

After you create any Virtual Channels, you must reset the RT Series DAQ Device
in order for them to take effect. The configuration information is downloaded to the
RT Series DAQ Device during the reset.

LabVIEW Real-Time 2-4 © National Instruments Corporation

 Configuring RT Series PXI Devices
Only a Host PC on
the same subnet can
initially configure
Network Settings.

Configure using the

Network Settings
Tab, click Apply, and
reboot the controller.

Select Help»Help
Topics»Remote
Systems for more
information.
ni.com

If you have an RT Series PXI Controller with the RT Engine installed, if you boot it up for the first
time after connecting it to the Network, you can see it from a Host PC on the same subnet with MAX.
If you have a non-RT Series PXI Controller, you can create a PXI Boot Disk from MAX (Tools»
Remote Systems»RT PXI Disk Utilities»Create PXI Boot Disk) which will allow you to boot the
controller into the RT Engine. (Only the PXI-8156B and PXI-8170 are supported at this time.) When
the PXI Controller is booted from the PXI Boot Disk for the first time, it creates the ni-rt.ini file
(which contains configuration information for the network settings and embedded version of LabVIEW)
and ni-rt folder on the c:\ drive. You can then configure the controller using MAX. For details on
using non-RT Series PXI Controllers with LabVIEW Real-Time, in MAX select Help»Help
Topics»Remote Systems»PXI Specific Information»PXI Controller with Windows Installed.
Configure the Network Settings, click Apply, and reboot the RT Series PXI Controller.
See the user manual for details.
Host Name refers to the name that the networked device will be known as on the network.
After an RT Series PXI Controller is configured, you can view it in MAX even if it is not on the same
subnet. To do this, right click on Remote Systems, select Create New, then select Remote Device
(not on the local subnet), click Next, then enter the IP address of the Remote System, and click on
Finish.
Other Functions:
• Reboot—this function allows you to remotely reboot the controller.
• Lock—you can protect the controller by providing a password.

© National Instruments Corporation 2-5 LabVIEW Real-Time

 Software for RT Series Devices

!Firmware is software that gets installed on

embedded hardware.
!Software for LabVIEW Real-Time consists of:
• Real-time operating system (RTOS).
• Embedded version of LabVIEW (the RT Engine).
• Device drivers.
!RT Series PXI—software is installed on hard
drive in c:\ni-rt directory.
!RT Series DAQ—software is downloaded
from Host PC to memory during reset.

ni.com

RT Series Networked Devices have firmware preinstalled; however, if it needs to be

reinstalled or upgraded, the LabVIEW RT Development System installed on your
Host PC contains the firmware images. Firmware can be reinstalled on RT Series
Networked Devices over the network using MAX.
The directory structure on RT Series Networked Devices is as follows:
c:\ Contains Ph_exec.exe (the RTOS). With the
PXI-8140 Series, the functionality of this executable
is contained in the BIOS.
c:\ni-rt\system Contains the RT Engine and device drivers.
c:\ni-rt\startup Initially empty. Stores applications.

LabVIEW Real-Time 2-6 © National Instruments Corporation

 Installing RT PXI Software

To install software, right-click in firmware window,

and select Install Firmware…
ni.com

If you wish to refresh the Remote Systems section of the Configuration window
in MAX, you can select View»Refresh, which refreshes the whole configuration
window, or you can select Tools»Remote Systems»Find Remote Systems, which
searches for Remote Systems on the same subnet as your Host PC.
Software images for RT Series Networked Devices are installed on the Host PC
in the c:\Program Files\National Instruments\RT Images
directory by default. Selecting Install Firmware… ,as shown above, installs
the appropriate software over the network on the RT Series Networked Device.
With LabVIEW 5.1.2, the Remote System Explorer is used to install firmware
on RT Series Networked Devices.

© National Instruments Corporation 2-7 LabVIEW Real-Time

 Remote DAQ Configuration
• From MAX: Tools»NI-DAQ Configuration»
Remote DAQ Configuration

• Enter Name or IP
address of RT Series
PXI Controller

• Clicking Save will send the configuration

to the RT Series Controller
ni.com

To configure the DAQ devices in the RT PXI Chassis, you will need to use Remote
DAQ Configuration. Through Remote DAQ Configuration, you can also configure
SCXI, virtual channels, and scales. When you finish the configuration, the
configuration file will be sent to the RT PXI Controller. You then need to reboot
the controller for some of the configuration changes to take effect.

LabVIEW Real-Time 2-8 © National Instruments Corporation

 RDA Server
The RDA Server runs on
RT Series PXI Controllers
by Default.

ni.com

To test DAQ devices on an RT Series PXI Controller, map the devices to

a Windows PC using MAX, then launch test panels on the RDA device.

© National Instruments Corporation 2-9 LabVIEW Real-Time

 Testing PXI DAQ Card Resources

Add Remote PXI Cards to Devices and

Interfaces on the Host PC (using RDA) to test
them (use Test Panels)

ni.com

The details of how to do this will be covered in an example.

LabVIEW Real-Time 2-10 © National Instruments Corporation

 Installing/Configuring FP-2000

FieldPoint Explorer 3.0 can be used for:

! Password Protect Resets
! Install/Upgrade Software
! Reboot Device
! View Installed Software
! Network Security
! Lock/Unlock System Configuration
! Device Location
! View Error Log

ni.com

Password Protect Resets—Determines whether the user is allowed to reset the

FieldPoint bank remotely. If enabled, the user must provide the password in order
to reset the system remotely.
Install/Upgrade Software—Use to install or upgrade LabVIEW RT software
on the selected FP-20XX.
Reboot Device—Use to reboot the selected FP-20XX.
View Installed Software—Use to view the LabVIEW RT and driver software versions
installed on the selected FP-20XX.
Network Security—You can limit host access to the FP-20XX by setting access
permissions for different host machines. The default setting allows Read/Write access
to all host machines.
Lock/Unlock System Configuration—Use to lock or unlock the system configuration
with a password.
Device Location—Use to change the search location for FP-20XX modules to the local
subnet or to a specified IP address.
View Error Log—Use to obtain detailed error information if you need to contact
National Instruments technical support.

© National Instruments Corporation 2-11 LabVIEW Real-Time

 Useful DOS Commands

!winipcfg or ipconfig
• Useful for finding IP configuration information
for the Host PC
!ping [IP Address]
• Useful for testing network connection
!ftp [IP Address]
• Useful for viewing RT Engine File System

ni.com

LabVIEW Real-Time 2-12 © National Instruments Corporation

 Exercise 2.1

Students become familiar with MAX in

relation to configuring RT Series Hardware.

Time to complete: 20 minutes

ni.com

© National Instruments Corporation 2-13 LabVIEW Real-Time

 Targeting RT Series Hardware
1. Select Operate»Switch Execution Target…
2. Select target, click OK.
LABVIEW
for
Windows
Target:
Host PC

3. Lower left
corner RT Development
indicates System
target. Target: RT Engine

ni.com

While developing a real-time application, LabVIEW Real-Time provides a user interface for
the RT Engine. This is done by “targeting” LabVIEW Real-Time to the RT Engine. When
LabVIEW Real-Time is targeted to the RT Engine, it is considered the RT Development
System. While developing RT applications in this manner, to the user, development is done the
same way as developing programs for Windows. Only, when the program is run, it actually runs
on the RT hardware and not on the windows machine.
To target LabVIEW Real-Time to your RT Series Hardware, start a new VI from LabVIEW
Real-Time, then:
1. Select Operate»Switch Execution Target.
2. Select the RT Engine you wish to target. In the target list, you will see an entry for every
RT Series DAQ Device configured in MAX. These are designated by DAQ::x RT
Engine on PCI-7030, where x is the device number in MAX. You will also see an entry
called RT Engine on Network. Select this if you wish to target LabVIEW Real-Time to
RT Series Networked Devices such as RT Series PXI Controllers. You will also see an
entry called Host PC (LabVIEW for Windows). Although this is actually not an RT
Engine, if you select this, then LabVIEW Real-Time will act as regular LabVIEW
for Windows.
3. While you are targeted, you will see the RT Engine indicated in the lower left corner
of both the front panel and diagram.

LabVIEW Real-Time 2-14 © National Instruments Corporation

 Target Selection Options
For PXI/PCI-7030 For Network Devices

• Reset Device—resets, • Machine Name/IP—for remote

installs firmware (~80 s) device
• Password—needed if remote
device is password protected

• Configure…—launches MAX
ni.com

When targeting LabVIEW Real-Time to an RT Engine on the network, such

as an RT Series PXI Controller, you must enter the RT Series PXI Controller’s
machine name or IP address as previously configured in MAX. If you provided
a password while configuring the network device, then you must also provide
it here.
When targeting an RT Series DAQ Device (plug in PCI or PXI board), if you are
targeting the board for the first time, then “Reset device” will be checked by default.
This is because it does not yet contain the software. During the reset, the software is
installed, which takes about 80 seconds. Each time the device loses power, it must
be reset.

© National Instruments Corporation 2-15 LabVIEW Real-Time

 Prompt for Target RT Engine
Tools»Options…»Miscellaneous

You can have the Target Selection Dialog Box

open automatically when LabVIEW is started by
selecting Prompt for Target Execution Engine.

ni.com

If you wish to have the Target Platform Selection Dialog Box appear automatically
when you first start LabVIEW, you can check the option as shown above.

LabVIEW Real-Time 2-16 © National Instruments Corporation

 Downloading a VI
• Download VIs by selecting:
Operate»Download Application
• VIs must be saved first.

• VIs are
automatically
downloaded when
the Run button
is pushed.

ni.com

When VIs are opened in the RT Development System, they are not downloaded to the
RT Engine until you manually download them.
You can see which VIs have been downloaded by selecting Browse»Show VI
Hierarchy. An inserted thumbtack indicates a downloaded VI. If the thumbtack is not
inserted, the VI is not downloaded.

© National Instruments Corporation 2-17 LabVIEW Real-Time

 Development State
Host PC RT Engine
LabVIEW
Compiled Code
for the VI

User
Interface
Thread
• UI Thread automatically updates interface
• Cannot run code targeted to Host PC
• Can run multiple VIs targeted to RT Engine
ni.com

A VI is in development state when it is open in LabVIEW Real-Time targeted to an RT

Engine (LabVIEW Real-Time will be in the RT Development System). What actually gets
downloaded to the RT Engine is the compiled code for the VI. When a VI is running in this
state, there is a thread (an independent sub-program) running on the RT Engine at normal
priority called the User Interface (UI) thread. The RT Engine UI thread exchanges
messages with the RT Development System to update the front panel indicators and also
read the front panel controls. The communication between the compiled code, the user
interface thread, and the Host PC is all transparent to the user. We will discuss threads,
multithreading, and priorities later in more detail.
While LabVIEW Real-Time is targeted to an RT Engine, any VI that you open and run will
run on the RT Engine. You cannot run any VIs targeted to the Host PC while LabVIEW
Real-Time is targeted to an RT Engine. This means that while targeted to an RT Engine,
you cannot perform any functions related to the Host PC such as File I/O, or accessing DAQ
devices in the Host PC.

LabVIEW Real-Time 2-18 © National Instruments Corporation

 Development State (Continued)

!When a VI is run, it is put into Run Mode.

To edit a VI after running, select
Operate»Change to Edit Mode
or press <CTRL+M>.
!All debugging features are supported in the
RT Development System (probes, execution
highlighting, single stepping, etc.), except for
the Call Chain ring.

ni.com

The Call Chain ring appears when a subVI is paused, and is not supported in the
RT Development System. (The Call Chain ring is also referred to in the Real-Time
User Manual, as is the Call List Window.)

© National Instruments Corporation 2-19 LabVIEW Real-Time

 Disconnecting From an RT Engine

!File»Exit
• Results in this dialog box:
!File»Exit without closing RT Engine VIs
• Automatically closes LabVIEW
• RT Engine VIs will continue to run
!Operate»Switch Execution Target
• RT Engine VIs will continue to run
!Deployed State – VIs are in Deployed State
after selecting Exit without closing RT
Engine VIs or Switch Execution Target

ni.com

If, while an RT Engine VI is running in the RT Development System, you exit

LabVIEW without closing RT Engine VIs, or Switch Execution Targets, then the
RT Engine VI continues to run, but instead of being in the development state, it is
in the deployed state. The UI thread on the RT Engine is no longer communicating
with the RT Development System.

LabVIEW Real-Time 2-20 © National Instruments Corporation

 Reconnecting to an RT Engine

!Front Panels not downloaded

!Reconnecting requires copy of VI on Host PC
!Changed or Missing VIs Dialog

ni.com

When you connect the RT Development System to an RT Engine with VIs already
downloaded or running, the RT Development will try to locate and open the
corresponding Front Panels on the Host PC. The Front Panel information is not
downloaded to the RT Engine and is only contained in the local copies of the VIs
on the Host PC. If the copy of the VI on the Host PC has been changed, then the
Changed or Missing VIs dialog box will appear. The user has the option of closing
the old RT Engine VIs and updating them with the new VIs, which will download the
new VIs, or the user can switch to Host LabVIEW without closing the RT Engine VIs.
If the VIs are missing, you will first be prompted to find them. If you are unable to
find the VIs, click cancel to make the Changed or Missing VIs dialog box appear.
Tip: You can FTP the Host PC’s copy of the VI to RT Engines that have storage
devices. This way, if the Host PC’s copy of the VI is lost or corrupted, you can
retrieve the backup copy from the RT Hardware using FTP.

© National Instruments Corporation 2-21 LabVIEW Real-Time

 Exercises 2.2 and 2.3

!Students use the RT Development System as

a user interface.
!Students use the Changed or Missing VIs
dialog box.

Time to complete: 20 minutes

ni.com

LabVIEW Real-Time 2-22 © National Instruments Corporation

 Deployed State
Host PC RT Engine
LabVIEW
Compiled Code
for the VI
co m
w ritten
r
Use User
Interface
E P ING
E
SLThread
• User written communication
• User Interface is now LabVIEW targeted
to Windows and designed by user
• User Interface Thread is dormant
ni.com

After the application has been developed, the RT Development System can be
disconnected from the RT Engine, and the RT Engine can run the application independent
of LabVIEW Real-Time on the Host PC. When an RT application is running independent
of the Host PC, it is considered to be deployed. LabVIEW Real-Time on the Host PC can
then be targeted to the Host PC itself. In this mode, LabVIEW Real-Time on the Host PC
is no longer running RT Development System, but is running as regular LabVIEW for
Windows. After an RT application is deployed, it can still communicate with programs
running in regular LabVIEW on the Host PC.
Since deployed applications running on RT Series Hardware do not have a User Interface
provided by the RT Development System, there are many cases in which you want to
create a companion application that runs on the Host PC or another PC. Your deployed
RT Engine VI can communicate with the companion VI. This companion application can
receive data from the RT Engine VI, analyze or log the data received, send data to the RT
Engine VI, and provide a way for a user to interact with your Deployed RT Engine VIs.
This is also true of applications that run on an RT Series DAQ device that need to have
data saved to disk. Since RT Series DAQ Devices (which plug into the PCI bus of the
Host PC) have no hard drive, they need to send that data to an application on the Host PC,
which can then save the data to the hard drive.
• VIs that run on the RT Series Hardware are referred to as RT Engine VIs.
• VIs that run on the Host PC are referred to as Host PC VIs.

© National Instruments Corporation 2-23 LabVIEW Real-Time

 Deployed State Communication
(Network)
Protocol Speed Loss? Ease
TCP Fast Lossless More Difficult
UDP Very Fast Lossy Moderate
DataSocket Fast Lossy* Easy
VI Server Slow Lossless** Moderate
*Data can be overwritten if items are not read from the DataSocket
Server before the next write.
**Not for large data transfers. Mostly used for monitoring, or controlling
one shot runs of remote VIs. VI Server is primarily suited for remote
control of VIs.

ni.com

It should be remembered that communication over a network is not deterministic.

LabVIEW Real-Time 2-24 © National Instruments Corporation

 Deployed State Communication
Networked Devices Networked
and RT Series DAQ Devices Devices only
TCP VI Server
UDP DataSocket

RT Series DAQ Devices only

Shared Memory

ni.com

It should be remembered that communication over a network is not deterministic.

Using DataSocket, networked RT Engine VIs can publish and read data to and from
a DataSocket Server running on another computer. Also, the RT Engine can
publish data to a DataSocket Server that many other machines can simultaneously
read from. LabVIEW Real-Time 5.1.2 RT Engines do not support DataSocket.
UDP is a faster communication protocol than TCP/IP, but it is not connection-based,
and data can be lost since there is no handshaking employed.
DataSocket and UDP are covered in more detail in the LabVIEW Advanced Course.

© National Instruments Corporation 2-25 LabVIEW Real-Time

 Deployed State Com. (Continued)
!Logos is supported on FieldPoint 2000
modules.
!CAN (Controller Area Network) is supported
in PXI for real-time two wire communication.
Can be synchronized with NI-DAQ
!Serial and GPIB communication are also
possible.

ni.com

LabVIEW Real-Time 2-26 © National Instruments Corporation

 TCP/IP

!TCP/IP functions the same between the

RT Engine and LabVIEW as between two
PC’s with LabVIEW.
!Covered in detail in the LabVIEW Advanced
Course.
!For RT Series DAQ Devices:
• TCP/IP communication only with LabVIEW on the
Host PC (no other computers)
• Address is DAQ::x, where x is the device number
in MAX
• The RT Engine cannot open a connection to the
Host PC
ni.com

TCP/IP is good for directly transferring large chunks of data between RT Engine VIs
and Host PC VIs.

© National Instruments Corporation 2-27 LabVIEW Real-Time

 TCP/IP Example

ni.com

The example above can be found in \Examples\comm\tcpex.llb.

LabVIEW Real-Time 2-28 © National Instruments Corporation

 VI Server

!VI Server functions the same as between the

RT Engine and LabVIEW as between two
PCs.
!Covered in detail in the LabVIEW Advanced
Course.
!For RT Series DAQ Devices:
• VI Server communication only with LabVIEW on
the Host PC (no other computers)
• Machine name is DAQ::x, where x is the device
number in MAX
• The RT Engine cannot open an application
reference to the Host PC
ni.com

One of VI Server’s capabilities is that it allows VIs to be called remotely as though

they existed on the caller’s system. VI Server has a lot of powerful functionality that
will not be covered in this course, but is covered in the LabVIEW Advanced Course.

© National Instruments Corporation 2-29 LabVIEW Real-Time

 VI Server Ex. (Call by Ref. Node)

ni.com

This is a simple example of using VI Server to run a VI on the RT Engine using

a Call by Reference Node. The VI being called using the Call by Reference Node
would reside on the RT Engine.

LabVIEW Real-Time 2-30 © National Instruments Corporation

 DataSocket
!Data is written to and read from a DataSocket
Server (The Server runs in Windows)
!Can have multiple clients
DataSocket Read
DataSocket Write

data data

ni.com

© National Instruments Corporation 2-31 LabVIEW Real-Time

 Network x.x.x.x Options…
!To set the options for networked RT devices,
after targeting the RT engine, select
Tools»Network x.x.x.x Options…
!Proper settings important for Network
Communication

ni.com

RT Engines on the Network have options that can be configured while targeted by
selecting Tools»Network x.x.x.x Options… See the RT Series PXI User Manual for
details on configuring these options.

LabVIEW Real-Time 2-32 © National Instruments Corporation

 Exercise 2.4

Students open and run a VI on the RT Engine

in the deployed state which communicates
with a corresponding Host PC VI using TCP.

Time to complete: 15 minutes

ni.com

© National Instruments Corporation 2-33 LabVIEW Real-Time

 RT Series DAQ Com (7030)

!TCP/IP and VI Server can be used with

PCI/PXI 7030 RT Series DAQ Devices.
This is implemented though a special area
of shared memory.
!1024 KB of shared memory can be accessed
directly using Shared Memory VIs.
!The rest of the chapter is optional if the class
does not need 7030 specific information.

ni.com

LabVIEW Real-Time 2-34 © National Instruments Corporation

 (7030) Shared Memory VIs
!1 KB (1024 bytes) of shared memory
accessible on RT Series DAQ Devices from
both the RT Engine and Host LabVIEW
!Common Shared Memory VI Inputs/Outputs:
• device
• byte offset
• device out
• next byte offset

ni.com

There is 1 KB of shared memory located on the RT Series DAQ Device processor

boards that can be accessed by both the RT Engine and the Host PC exclusively
using Shared Memory VIs. TCP/IP and VI Server also use shared memory to
communicate with the Host PC, but it is not accessible by the user, and does not
include the 1 KB reserved for the Shared Memory VIs. There is a pallet of VIs
designed exclusively for writing and reading values to and from shared memory.
From the Functions pallet select RT»RT Series DAQ.
Many of these VIs share common inputs and outputs such as device and byte offset.
For programs running on the Host PC, the device number is the device number for
the processor board (i.e., 7030) found in MAX. For programs running on the RT
Engine, the device number is 0. The byte offset is the location of the beginning byte
for the data you will write or read. Since there is one kilobyte of shared memory, the
byte offset can be a value from 0 to 1023. If you try to read or write beyond the
1 KB of shared memory, you will get an error. Device out is a duplicate of device
input. Another commonly used output is next byte offset which is the next available
location after this VI runs.
The shared memory VIs are typically used to pass parameters or variables rather
than large chunks of data. TCP/IP is better suited for large data transfers.

© National Instruments Corporation 2-35 LabVIEW Real-Time

 (7030) RT Peek Poke with Error
!VIs read and write directly to Shared Memory,
with error checking
!Based on RT Low Level Peek Poke VIs
!No guarantee data written will be read
(you must write your own handshaking)
!Example - writing and reading int8
and double:
RT Engine Host PC

ni.com

If a peek is done at the same time as a poke, some bytes may contain old data while
others contain new data, which results in a corrupt value.

LabVIEW Real-Time 2-36 © National Instruments Corporation

 (7030) RT Shared Memory R/W VIs
!Similar to Peek Poke VIs, but use a flag
to determine if data is new or old.
!Data will not be written if it is the same as the
previous data, or if the data at that byte offset
is still new.
!Data is not read if data flag shows data is old,
and previous data input will be returned.

ni.com

If a peek is done at the same time as a poke, some bytes may contain old data while
others contain new data, which results in a corrupt value. The Read/Write VIs avoid
this problem by not writing a new value until the entire old value has been read.
The tip strips on the diagrams show that the highlighted input is for “previous data
(DBL).”

© National Instruments Corporation 2-37 LabVIEW Real-Time

 (7030) RT Board Utilities
•To find size of shared memory •To control LEDs

•To Pass Errors back from the RT Engine to the Host PC

Use on the RT Engine

Use on the Host PC
•To Write and Read arrays (writes/reads verification number
before and after array to ensure data is not corrupt)

ni.com

Use the RT Board Utilities VIs located on the Functions»RT»RT Series DAQ»
RT Board Utilities palette to check the shared memory size, toggle the board LEDs,
and facilitate high-level communication between the RT Series DAQ Device and
the host PC.

LabVIEW Real-Time 2-38 © National Instruments Corporation

 (7030) Incremental Single VIs
Rather than read or write many parameters
each loop iteration, these VIs write or read a
portion of the data array each iteration.
• Reads only 2 bytes of
the array each time it is
called

• Writes only 2 bytes of

the array each time it is
called

ni.com

When a large number of parameters needs to be passed between the RT Engine and
the Host PC, the Incremental Single Read and Incremental Single Write VIs are
very useful. Rather than transferring all of the parameters each iteration, which can
slow down your loops, these VIs allow you to transfer a small portion of the
parameters in each loop iteration. Only a portion of the data array is written or read
from shared memory each time. Although this allows your loops to run faster, this
means that it may take several loop iterations for a parameter to be transferred, and
sometimes parameters values are not transferred at all, since they may change many
times before actually being transferred.
There is a concept document in the National Instruments Developer Zone called
“Incremental Single Read and Incremental Single Write VIs” which discusses these
VIs in more detail.

© National Instruments Corporation 2-39 LabVIEW Real-Time

 Exercises 2.5 and 2.6

Students use shared memory VIs

(for RT Series DAQ Devices only).

Time to complete: 20 minutes

ni.com

LabVIEW Real-Time 2-40 © National Instruments Corporation

 Chapter 2—Summary

!Use MAX for configuring RT Series Devices

!LabVIEW Real-Time is used to develop
VIs on the Host PC which are downloaded
to the RT Engine
!Development State vs. Deployed State
!User Interface Thread runs while in
development state at normal priority
!User must write communication for VIs
in deployed state

ni.com

© National Instruments Corporation 2-41 LabVIEW Real-Time

 Exercise 2-1
Objective: To become familiar with MAX in relation to configuring
RT Series Hardware.

This exercise assumes that the RT Series Hardware that you will use has been
previously installed and configured.

1. Launch MAX. To do this, select Start»Programs»National Instruments»

Measurement & Automation Explorer.
2. Refresh MAX by selecting View»Refresh, or press <F5>.

PART 1. Finding Device numbers for RT Series DAQ Device processor boards and
their daughter cards.
1. Expand My System, and then Devices and Interfaces if they are not already
expanded. Do this by double clicking on the item you would like to expand.
2. Any RT Series DAQ Devices in your PC will show up here. Double click on
the RT Series DAQ Device (if you have one installed, such as a PCI-7030-6XXX)
to expand it to see the DAQ daughter card. Write down the device numbers
of the processor board and the DAQ daughterboard. (If you do not have
an RT Series DAQ Device, skip to Part 2.)

RT Series DAQ Device:

Processor Board Device Number _________
Daughterboard Device Number _________

3. Test the resources for each RT Series DAQ Device by right clicking on the
processor board and selecting Properties…»Test Resources. Notify the
instructor if your device fails the test.

Tested Passed? " Yes " No

Note: You can also configure SCXI Chassis, Virtual Channels, and Scales for the
RT Series DAQ Device’s daughter card just like you would for normal DAQ cards
in MAX, however, this is beyond the scope of this class. You cannot test daughter
cards or virtual channels for daughter cards from MAX.

PART 2. Using MAX to view configuration information for RT Series PXI Controllers.
1. Select Tools»Remote Systems»Find Remote Systems. This will locate any
RT Series PXI Controllers on the same subnet as your Host PC. They will then be
listed in the Configuration window of MAX under My System»
Remote Systems.
2. Expand Remote Systems if it is not already expanded.
3. If there are any RT Series PXI Controllers on the same subnet as your Host PC,
they will appear here. Click on the name of the RT Series PXI Controller that you
will be using (consult the instructor about which controller you will use if there

LabVIEW Real-Time 2-42  National Instruments Corporation

 are multiple controllers listed). If you will not be using a controller, just read
through the remaining steps.
4. Select the Network Settings tab and write down the IP address.

IP Address _____ . _____ . _____ . _____

5. Select the Firmware tab and observe the versions installed on the RT Series
PXI Controller.
6. From MAX, select Tools»NI-DAQ Configuration»Remote DAQ
Configuration…, enter the IP Address and click OK. Expand the NI-DAQ
Devices section. Write down the Device Numbers for each of the PXI DAQ cards.

Device ___: PXI-__________

Device ___: PXI-__________
Device ___: PXI-__________
Device ___: PXI-__________

Note: You can also configure SCXI Chassis, Virtual Channels, and Scales for the
RT Series Controller using Remote DAQ Configuration, however, this is beyond
the scope of this class.

7. Click Exit when you are finished.

PART 3. Using MAX and RDA to test PXI DAQ board resources in PXI chassis with
RT Series PXI Controllers. (Skip this part if you are using an RT Series DAQ Device.)
1. In MAX, right-click on Devices and Interfaces and select Create New…
2. Select RDA/Ethernet Device and click Finish.

 National Instruments Corporation 2-43 LabVIEW Real-Time

 3. In the Select Remote Computer dialog screen that appears, enter the Remote
Computer Name/IP Address for your PXI controller then click Next.

4. Select the device you want to test and click Next.

LabVIEW Real-Time 2-44  National Instruments Corporation

 5. Select the Device Number that the card will be listed under for your local
machine, then click Finish. The device number for the card on your local machine
is different from the device number of the card on the remote system, although
they can be the same.

6. The card should now be listed under Devices and Interfaces. Test the resources
for the remote DAQ board by right-clicking on it and selecting Properties…»
Test Resources. If the device fails the test, reboot the RT Series Controller and
try the test again. Notify the instructor if the device fails the test a second time.

Tested Passed? " Yes " No

 National Instruments Corporation 2-45 LabVIEW Real-Time

 7. Right-click on the device and choose Test Panels. Test the DAQ channels
to verify the device is working properly.

8. After testing the DAQ card, you can delete it from Devices and Interfaces.

Exercise 2-2
Objectives: 1) Configure LabVIEW Real-Time to prompt for execution each
time LabVIEW Real-Time is started. 2) Run a VI on the target platform
using the RT Development System as your user interface.

1. Start LabVIEW Real-Time. From a new VI, select Tools»Options…

2. Select Miscellaneous from the top pull-down menu.
3. Check the box in front of Prompt for Target Execution Engine* and click OK.
4. Close LabVIEW Real-Time and then start it again. This time the Select Target
Platform dialog box will automatically appear.
5. From the Select Target Platform dialog box, select your RT Engine. Click OK.
6. Open the Tank Simulation VI located in c:\Program Files\National
Instruments\LabVIEW 6\Examples\Apps\tankmntr\Tank
Simulation.vi.
7. Open the VI Hierarchy window by selecting Browse»Show VI Hierarchy and
notice that the thumbtack for Tank Simulation.vi is in the horizontal
position, which indicates that it has not been downloaded.
8. Download the VI by selecting Operate»Download Application. In the
Hierarchy Window, select View»Redraw. Notice in the VI Hierarchy Window
that now the thumbtack is in the vertical position, which indicates that it has now
been downloaded.

LabVIEW Real-Time 2-46  National Instruments Corporation

 9. Click the Run button. The RT Development System is now providing a user
interface for the RT Engine on the target platform. Experiment with the control
values to control the Tank Simulation VI.
10. Exit LabVIEW Real-Time by selecting File»Exit Without Closing RT Engine
VIs. This exits the RT Development System but leaves the embedded Tank
Simulation VI running on the target platform.
11. Start LabVIEW Real-Time again. From the Select Target Platform dialog box,
select the same target platform. Clear the Reset checkbox if you target an RT
Series DAQ device. Click OK. Because you chose to exit without closing the
embedded VIs, LabVIEW Real-Time establishes communication with the target
platform, opens the Tank Simulation VI in the RT Development System
on the Host PC, and provides the RT Development System with the most recently
updated data from the embedded LabVIEW Real-Time VI still running on the
target platform.
12. Exit LabVIEW Real-Time by selecting File»Exit. Close all RT Engine VIs
and Quit. Do not save changes. This closes LabVIEW Real-Time and stops and
closes the Tank Simulation VI on the target platform.

Exercise 2-3
Objective: To see how an embedded VI interacts with a copy of the VI
on the Host PC. You will use the Changed or Missing VIs dialog box.

1. Start LabVIEW Real-Time and select your RT Engine from the Select Target
Platform dialog box.
2. Select File»Open and open the TCP - RT Engine VI from c:\Program
Files\National Instruments\LabVIEW 6\Examples\RT\RT
Communication.llb.
3. Select File»Save As and save the VI to your desktop.
4. Download the VI by selecting Operate»Download Application. Do not run the VI.
5. Select File»Exit Without Closing RT Engine VIs.
6. Delete the VI you just saved to the desktop.
7. Launch LabVIEW Real-Time and select the appropriate target platform from
the Select Target Platform dialog box. Clear the Reset checkbox if you target
an RT Series DAQ device. Because the RT Development System cannot find
the local copy of the VI, it prompts you to locate it. You could select the
TCP - RT Engine VI from Examples\RT\RT Communication.llb, but
for this exercise, click Cancel. The Changed or Missing VIs dialog box appears.
When the RT Development System detects that there is an embedded VI on the
target platform, it attempts to open the front panel from the local copy of the VI
on the host PC hard drive. If the RT Development System detects a change in
version or cannot locate the host copy of the VI, the Changed or Missing VI
dialog box appears. When the Changed or Missing VI dialog box appears you
have the option to close all the embedded VIs and update them with the copy
on the host PC or to exit the RT Development System and not update the
embedded VIs.

 National Instruments Corporation 2-47 LabVIEW Real-Time

 8. Click the Close all RT Engine VIs and Update button in the Changed or
Missing VIs dialog box.
9. Repeat steps 2 through 8, except in Step 6, instead of deleting the VI, add an
indicator to the front panel and save the VI. To do this, open the VI in LabVIEW
targeted to the Host PC, and when finished close LabVIEW. In Step 7,
LabVIEW Real-Time will not prompt you for the VI’s location, but will
immediately display the Changed or Missing VIs dialog box. Click the Close
all RT Engine VIs and Update button in the Changed or Missing VIs dialog
box. This will download the new VI.
10. Exit LabVIEW Real-Time by selecting File»Exit and close all RT Engine VIs.

Exercise 2-4
Objective: To use a VI in LabVIEW targeted to the Host PC to serve
as a User Interface for a deployed VI.

This example uses an RT Engine VI and a Host PC VI. The RT Engine VI runs on the
target platform in deployed state and communicates with the Host PC VI that runs on
the Host PC. The Host PC VI provides the user interface.

1. Start LabVIEW Real-Time.

2. From the Select Target Platform dialog box, select your RT Series device.
Click OK.
3. Open the TCP - RT Engine VI, found in c:\Program
Files\National Instruments\LabVIEW 6\Examples\RT\RT
Communication.llb and click the Run button. The RT Development
System automatically downloads the VI to the target platform, which runs the VI.
4. Select Operate»Switch Execution Target. This disconnects the RT
Development System from the target platform but leaves the TCP - RT Engine VI
running on the RT Engine in deployed state.
5. In the Select Target Platform dialog box, select Host PC (LabVIEW for
Windows) to use LabVIEW on Windows to run VIs rather than the RT Engine on
the RT Series device. Click OK.
6. Open the TCP – Host PC VI, found in c:\Program Files\National
Instruments\LabVIEW 6\Examples\RT\RT Communication.llb.
7. On the TCP – Host PC VI front panel, you must change the machine name/IP
address control to the address of the RT Series device. For RT Series DAQ
devices, use DAQ::x, where x is the device number of the desired RT Series
DAQ device on your system. For networked RT Series devices, use the IP address
of your networked RT Series device.
8. Click the Run button. Because you selected the Host PC rather than the RT
Engine in the Select Target Platform dialog box, the TCP - Host PC VI
runs on the Host PC. It receives and displays data through TCP/IP from the
TCP - RT Engine VI running on the RT Series device.
9. Change the function control from random to sine wave. The desired function
type is sent through TCP/IP to the RT Engine VI, which then changes the function
it generates and sends to the Host PC.

LabVIEW Real-Time 2-48  National Instruments Corporation

 10. Select Operate»Switch Execution Target and select your RT Engine.
Select File»Exit to exit LabVIEW Real-Time. Do not save your changes.
Exercise 2-5
Objective: To use a VI running on the Host PC to communicate with
a deployed VI (also called embedded VI) using shared memory VIs.
This example is for use with RT Series DAQ Devices only!

1. Start LabVIEW Real-Time.

2. From the Select Target Platform dialog box, select the appropriate RT Series
DAQ board from the Select Target Platform dialog box. Click OK.
3. Open the Peek Poke I32 - RT Engine VI, found in the c:\Program
Files\National Instruments\LabVIEW 6\Examples\
RT\RTCommunication.llb library and examine the diagram. Run the VI.
4. Select Operate»Switch Execution Target to close the RT Development System
but not the embedded VI.
5. Select Host PC (LabVIEW for Windows) from the Select Target Platform
dialog box. Open the Peek Poke I32 - Host PC VI, found in the c:\Program
Files\National Instruments\LabVIEW 6\Examples\RT\
RTCommunication.llb library and examine the diagram. Refer to
Windows»Show VI Info of these VIs for more information. Run the VI.

Tip: Although peek and poke VIs are the low level, fastest form of communication,
using several of these VIs in a control loop can slow down your application. If you
need a loop with communication to run at maximum rates, use the RT Incremental
Single Write VI and RT Incremental Single Read VI, found on the RT»RT Series
DAQ»RT Board Utilities palette.

The size of the shared memory is limited to 1 KB. If you need to transfer several
megabytes of data, you must separate the data into smaller portions and then transfer
them. In doing so, you must make sure that you do not overwrite data in the shared
memory before you read it. TCP/IP VIs manage flow control and are more convenient
for bulk transfers.

Exercise 2-6
Objective: To use Shared Memory VIs to communicate between the Host
PC and the RT Engine running on an RT Series DAQ Device. This example
is not necessarily a use case, but rather an exercise that teaches the
concept of how some of the shared memory VIs work.

1. Build the following VI for the Host PC.

• This VI will write an array of 10 elements to shared memory using the
RT write SGL array VI in each iteration. This array will be written
to byte offsets 0 through 43 of the RT Series DAQ Device’s shared memory.
• It will then read an array that is written by the RT Engine to byte offsets 44
through 103 using the Incremental Single Read VI. The Incremental Single
Read VI does not read the array all at once. Instead, each main loop iteration it

 National Instruments Corporation 2-49 LabVIEW Real-Time

 checks a flag in front of one of the array elements in shared memory. If the
flag indicates new data (flag =1), then over the next 2 main loop iterations,
it will read the element, and then in the next main loop iteration it will change
the flag to a zero, indicating that the data is now old.
• At the end of each main loop iteration, it will also check for an error message
from the RT Engine VI starting at byte offset 104.
• The shared memory VIs are written so that the next byte offset is calculated
automatically, so it does not need to be specified explicitly.
• Change the representation Element data input of the initialize array to single
precision.
• Change the device number to that of your RT Series DAQ Device.
• Save the VI when you are finished.

LabVIEW Real-Time 2-50  National Instruments Corporation

 2. Build the following VI. This VI will read an entire array of 10 elements from
shared memory each loop iteration and write that array back incrementally
(only a small amount of the array will be written in each iteration) to a different
location in shared memory.
• The VI first initializes the shared memory to zero.
• In each main loop iteration, the VI reads an array of 10 elements from shared
memory using RT read SGL array VI.
• In each main loop iteration, the Incremental Single Write VI will be
performing one of several operations, depending on what it did on the last
iteration:
− Check if one of that data input elements is new or not. If it is not new,
do nothing. If it is new, check the shared memory flag to see if the
previous data element written to shared memory was read yet.
− Write the first half of the single precision element (two bytes).
− Write the second half of the single precision element (two bytes).
− Write a 1 to the flag location indicating an array element in shared memory
is new (two bytes).
• Save and run the VI targeted to the RT Series DAQ Device.

 National Instruments Corporation 2-51 LabVIEW Real-Time

 3. Switch the execution target to the Host PC and run the VI created in Step 1.
4. Change the element data and notice how the data gets read back in increments.
When trying to achieve faster loop rates, writing whole arrays each iteration may
hinder your loop rates. This is where the Incremental Single Read/Write VIs help.
Instead of writing a whole array each iteration, they only write or read a portion of
the array. The trade off is that some of the elements of the array may not be
communicated right away, and some values may be missed entirely.
If you are only monitoring values, or communicating parameters that change
slowly, then the Incremental Single Read/Write VIs will allow you to do this
while maintaining faster loop rates.
5. Exit LabVIEW.

LabVIEW Real-Time 2-52  National Instruments Corporation

 Lesson 3

Multithreading and Priorities

You Will Learn:

A. About multithreading and why it is useful in
LabVIEW RT applications
B. About VI priorities
C. Inter-thread communication with global
variables

ni.com

© National Instruments Corporation 3-1 LabVIEW Real-Time

 What Is Multithreading?

!Thread—an independent sub-program that

runs within a program
!Multithreading—the ability to break your
program into multiple threads of varying
priority
!Priority—threads can be given different
priority levels. Higher priority threads preempt
lower priority threads.

ni.com

A thread is an independent sub-program that runs within a program or process.

LabVIEW is a process.

LabVIEW Real-Time 3-2 © National Instruments Corporation

 Multithreading Metaphor
Time Critical Normal
Priority (one VI only) Priority

Tasks

Receptionist (OS) One Mechanic

(processor)

ni.com

MECHANIC = PROCESSOR: In this example there is only one mechanic (Processor)

which performs the repairs (processing).
RECEPTIONIST = OPERATING SYSTEM: The receptionist (Operating System)
lines up the service requests as they arrive. He will schedule them in order of arrival,
unless a higher priority customer arrives, in which case the receptionist will put that
customer before the lower priority customers.
ONE TOWN AMBULANCE = ONE TIME CRITICAL PRIORITY VI ONLY :
In this example, there is one automobile in the town which has the highest priority:
the ambulance. This is recommended for Real-Time programs as well.
WORK ON CARS OF EQUAL PRIORITY AT THE SAME TIME = MULTI-
TASKING WITHIN EQUAL PRIORITIES: The mechanic is able to work on multiple
cars of the same priority at the same time, and progress them all to completion,
working a little bit on one car, and a little on the next. (Equal priority threads share
the CPU.)
PREEMPTION: If the mechanic is working on lower priority cars, and a car of higher
priority arrives, he will put the lower priority cars aside to work on the higher priority
cars. When he is finished with the higher priority cars, he can return to the lower
priority cars.
NEVER FINISH LOW PRIORITY CARS = STARVATION: If there always are
higher priority cars that need work, then lower priority cars will never get worked on.

© National Instruments Corporation 3-3 LabVIEW Real-Time

 Scheduling Threads

!Round Robin
• Threads of equal priority
receive equal time slices
• Might take several turns for
a thread to complete

!Preemptive
• Higher priority thread
immediately suspends
execution of all lower priority
threads

ni.com

LabVIEW Real-Time’s real-time operating system (RTOS) implements a combination

of round robin and preemptive thread scheduling.
Round robin scheduling applies to threads of equal priority. Equal shares of CPU time
are allocated between the equal priority threads. For example, each normal priority
thread is allotted 10 milliseconds to run. It executes all the tasks it can in 10 milliseconds
and whatever is incomplete at the end of that period must wait to complete during the
next allocation of time.
Conversely, preemptive scheduling means that any higher priority thread that needs to
execute will immediately suspend execution of all lower priority threads, and will begin
to execute.

LabVIEW Real-Time 3-4 © National Instruments Corporation

 Why Multithreading?

!We need to differentiate between time-critical

tasks and non-critical tasks.
• Some time-critical tasks:
– Control Loop
– Safety Monitoring
• Some non-critical tasks:
– Communication
– Data logging
!For Real-Time performance, an OS that can
give scheduling priority to time-critical tasks is
needed.

ni.com

Multithreading is useful when you have parts of your code that do not need to be
deterministic, or if parts of your code rely on non-deterministic I/O. For example,
putting network communication tasks inside the time-critical loop may harm
determinism. If you make your time-critical code rely on responses from another
PC over the network, then if the other PC does not reply in time, it may cause you
to miss a deadline. In order to remedy this, you can break the threads up into time-
critical tasks and tasks that are not time-critical, then you can assign higher priority
to time-critical tasks to ensure that they always finish on time.

© National Instruments Corporation 3-5 LabVIEW Real-Time

 Multithreaded RT Application

Not Time Critical

Time Critical
(Also not
deterministic)

•Communication with Host

•File I/O to hard drive
•Serial Communication
•NI-DAQ •GPIB Communication
•NI-CAN
•NI-VISA with PXI
•NI-Motion

ni.com

LabVIEW Real-Time 3-6 © National Instruments Corporation

 Time Critical Priority

!To achieve “hard real-time” determinism is to

have the VI priority set to time critical.
In this priority on the RT Engine, no parallelism is
allowed, thus no parallel loops. This is to ensure
deterministic behavior.
!Only one VI should be set to time-critical
priority (with one loop). Nested loops are
allowed.

ni.com

© National Instruments Corporation 3-7 LabVIEW Real-Time

 Setting a VI’s Priority
1. Right Click on Icon»VI Properties…
2. Select Execution Category
3. Select Priority Level

• SubVIs inherit the

highest priority of any
of its callers in memory
(only higher priorities
inherited, not lower)
ni.com

To set the priority of a VI, right-click on the icon and select VI Properties…,
then select the Execution Category and then the Priority Level. The highest priority
level that allows multithreading is time-critical priority. Subroutine priority does not
allow multithreading, and will be discussed later.
SubVIs will inherit the priority of any of its callers in memory. This inheritance
property allows for upward priority inheritance only. For example, even if a normal
priority VI calls a normal priority subVI, if that subVI is in a time-critical priority VI
that is in memory, then the subVI will run at time-critical priority, even though
a normal priority VI called it.

LabVIEW Real-Time 3-8 © National Instruments Corporation

 Sleeping

!Sleeping is the act of suspending the

execution of a VI or thread.
!VIs can go to sleep, allowing lower priority
VIs to run.
!There are many ways to cause a VI to sleep,
the first of which is:

More ways discussed later…

ni.com

Sleep mode can be thought of as a programmatic tool that a VI can use to proactively
take itself off of the LabVIEW and operating system scheduling mechanisms. Users
can programmatically provide a sleep mode in a VI by using the functions Wait Until
Next ms Multiple, or Wait (ms).

© National Instruments Corporation 3-9 LabVIEW Real-Time

 Sleeping and Time Critical Priority

!If any VI in a time-critical thread goes to

sleep, the entire thread goes to sleep –
(Feature unique to LabVIEW Real-Time)

!Parallel loops not recommended in

time-critical VIs—multi-tasking is disabled
(Feature unique to LabVIEW Real-Time)

ni.com

LabVIEW Real-Time’s time-critical priority threads have a unique characteristic that

is different from normal LabVIEW scheduling. If any VI on a time-critical thread goes
to sleep, the entire thread goes to sleep. Even other VIs on the same thread that did not
request to go to sleep will sleep. This is only the case for the time-critical priority
(highest) setting. Conversely, if two VIs (or two loops for that matter), are executing on
the same thread (other than time-critical priority), and one of them goes to sleep, the other
VIs on the same thread will continue to execute. In other words, LabVIEW Real-Time’s
execution system will not schedule time-critical priority operations from parallel VIs,
or loops, when any one of them sleeps in the same time-critical priority thread. All other
priority threads in LabVIEW Real-Time, and all threads in normal LabVIEW, will
continue to schedule operations from parallel loops, and/or VIs, in similar threads.
Given the cooperative multitasking nature of scheduling multiple time-critical priority
(highest) threads, it is recommended that only one VI, and loop, ever be used with the
time-critical priority (highest) setting. This is the only way to receive a guarantee in
deterministic execution.

LabVIEW Real-Time 3-10 © National Instruments Corporation

 Sleep Mode (Continued)

!Millisecond resolution is available using Wait

Until Next ms Multiple (loop rates of 1000 Hz,
500 Hz, 333 Hz, 250 Hz...).
!What if we want a smaller sleep resolution,
i.e., on the order of microseconds or
nanoseconds?
We will see later that we can achieve that using
hardware tools.

ni.com

When controlling the rate of a software loop by using the Wait Until Next ms
Multiple function users can only achieve rates in 1 millisecond multiples. This means
you can either run your loop at full speed providing no sleep at all, or you can achieve
loop rates of 1000 Hz, 500 Hz, 333.33 Hz, 250 Hz, 200 Hz, etc. We will discuss
a method of implementing loop rates with a finer sleep mode resolution than
1 millisecond multiples later in this presentation.

© National Instruments Corporation 3-11 LabVIEW Real-Time

 Starvation

Higher Priority Process A blocks out lower

process
Time slice Process A Process B Process C
(Time (Above (Normal
Critical) Normal Priority)
Priority)
1 Running Ready Ready
2 Running Ready Ready
3 Sleeping Running Ready
4 Running Ready Ready

ni.com

To understand starvation, consider the following example:

Three processes compete for processor or any other resource. Since process A has
a time-critical priority, it will run until it is done using that resource. At that point,
one of the two other processes, say Process B, will be allowed to use the resource for
one time slice. If at that point process A is ready to run; it will again take control of the
resource and run until it has finished its critical section. Again, one of the two processes
will be allowed to run. Once again, process B will be chosen. Process C can be blocked
from the resource indefinitely.
If process A is not put to sleep for enough time for both processes to run, it can happen
that a lower priority process is never allowed to run.

LabVIEW Real-Time 3-12 © National Instruments Corporation

 User Interface Thread

!Runs at Normal Priority

RT Engine
!Used to communicate LabVIEW
with the RT Development Compiled Code
System for the VI
!Is sleeping while the VI is
in the Deployed State
!Can be starved if a higher User
priority VI sleeps for very Interface
little or no time Thread

ni.com

When in the development state, the compiled code will only run in real-time
if there is no interaction with the front panel.

© National Instruments Corporation 3-13 LabVIEW Real-Time

 Exercise 3.1

Students observe the effects of changing the

priority of a VI.

Time to complete: 15 minutes

ni.com

LabVIEW Real-Time 3-14 © National Instruments Corporation

 Inter-thread Communication
Global variables are one tool for
communicating between threads
• Can cause jitter (discussed later)
• Lossy—a global variable may be
written to many times before being
read

Time Critical Other Priority

? •Network Com
•File I/O
•Etc.

ni.com

Global variables pass information between locations in your application that you
cannot connect with a wire. You can use global variables to access and pass data
among several VIs.
Global variables are built-in LabVIEW objects. When you create a global variable,
LabVIEW automatically creates a special global VI, which has a front panel but no
block diagram. Add controls and indicators to the front panel of the global VI to
define the data types of the global variables it contains. In effect, this front panel
is a container from which several VIs can access data.
Global variables can add jitter to the program because they are a shared resource.
Shared resources and better methods of inter-thread communication are discussed
later.

© National Instruments Corporation 3-15 LabVIEW Real-Time

 Multithreaded Communication

Deployed RT Engine Host PC

Application
Normal Priority Network
Comm. VI

Inter-thread
Communication

Time Critical
VI

ni.com

Here is an example of a simple distributed application with multithreading. The RT

Engine is in the Deployed State. It has two loops, a time-critical loop and a normal
priority loop. This can be achieved by having the top level VI set to normal priority,
with a communication loop inside the normal priority VI. The time-critical loop can
be inside a subVI of the normal priority VI. This subVI should be set to time-critical
priority. The time-critical VI should execute independently of the normal priority VI.
The Normal Priority Communication loop can have some type of communication
method in it like TCP/IP, UDP, or DataSocket.
In this architecture, we separate all the tasks that are critical for the application from
non-critical ones. For example, control loops and critical event responses should all
be in a separate time-critical loop. All the tasks that are less critical or tasks that can
jeopardize determinism such as communication back to the host or logging to file,
should occur in a separate normal priority loop.

LabVIEW Real-Time 3-16 © National Instruments Corporation

 Inter-thread com. with Global

Deployed RT Engine Host PC

Application
Network
Normal Priority VI
VI

Inter-thread
Global
Communication

Time Critical
VI

ni.com

You can replace the normal priority communication loop with a VI that just reads
and writes to the global variables which the time-critical VI also reads and writes to.
The Host PC could then call the normal priority VI through VI Server to get the
variable values. This form of communication is lossy, though, since the global
variables could be updated many times before the normal priority VI gets called
by the Host PC. There are many examples of this type of communication in the
Examples\RT folder.

© National Instruments Corporation 3-17 LabVIEW Real-Time

 Multithreading Example

Deployed RT Engine Host PC

Application
Communication.vi
Network

Inter-thread
Communication
with Globals

Wait Until Next ms

Multiple used to allow
lower priorities to run

Time Critical.vi
ni.com

The simplest way to assure that starvation does not occur is to add Wait VI’s to your
loops. This will put your high-priority thread to sleep and will give the other threads
processor time. You want to make your sleep time long enough for the other threads
to accomplish their tasks, but short enough that your determinism is not compromised.

LabVIEW Real-Time 3-18 © National Instruments Corporation

 Multithreaded Data Logging

Deployed RT Engine Perform file reads

Application and writes in 512
byte packets for
Normal Priority optimal speed.
File I/O VI

Inter-thread
Communication

Time-Critical
VI

ni.com

It is important to do all file I/O operations in a normal priority loop. This is because file I/O
is inherently non-deterministic. File I/O calls in the time-critical loop can jeopardize the
determinism of the entire program.
When transferring data to disk, the transfer rate is largely dependent on the size of the read
or write you are performing. In reading and writing, the optimal size is 512 bytes. This is
because 512 bytes is the size of a sector on a hard disk, and is the basic unit that is read or
written from a disk. If you wanted to read 32 bytes, you would have to read an entire 512
byte sector first, then pull out the 32 bytes that you need from it. The same issue arises with
writing—if you want to write 32 bytes, you have to read the full sector first, then replace
the bytes, then write the entire sector back to disk. Since LabVIEW Real-Time does not
do disk caching on the embedded side, we see these effects clearly. On platforms with disk
caching, the performance is typically higher and more balanced at the expense of
determinism.

© National Instruments Corporation 3-19 LabVIEW Real-Time

 Exercise 3.2

Students use global variables

to communicate between threads and
the VI Server Call by Reference Node
to communicate between the RT Engine
and Host PC.

Time to complete: 20 minutes

ni.com

LabVIEW Real-Time 3-20 © National Instruments Corporation

 Exercises 3.3 and 3.4 (Optional)

Students build multithreaded RT Engine VIs

that communicate with the Host PC VI using:
• DataSocket—Exercise 3.3
• VI Server Get/Set Control Value—Exercise 3.4
(Optional)

Time to complete: 40 minutes

ni.com

© National Instruments Corporation 3-21 LabVIEW Real-Time

 Chapter 3—Summary

!LabVIEW uses a combination of round robin

and preemptive scheduling
!Time-critical tasks need to be in a separate
thread from rest of tasks
!Time-critical threads should have some sleep
time to avoid starvation of lower priority
threads
!Only one time-critical VI with no parallel loops
should be used

ni.com

LabVIEW Real-Time 3-22 © National Instruments Corporation

 Exercise 3-1
Objective: To observe the effects of changing the priority setting of a VI.

1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. For RT Series DAQ Devices, select the Reset checkbox to reboot the hardware
if needed.
3. Click the OK button.
4. Open the Priority Trouble VI from c:\Program
Files\National Instruments\LabVIEW 6\Examples\RT\RT
Tutorial.llb.
5. Right-click on the VI icon on the front panel and select VI Properties…
6. Select Execution from the pull-down menu. Notice that the priority is set to
time- critical priority (highest). This makes the VI run in the highest priority
thread on the RT Engine.
7. On the front panel, set the msec to wait control to greater than 0. 10 is the default.
Run the VI. The chart displays a sine wave.
8. Slowly decrement the msec to wait control. With each step, the chart display
speeds up. However, when the msec to wait control reaches zero, the front panel
no longer updates.

The reason for this behavior is that the RT Engine can no longer communicate
with the RT Development System front panel. Because the time-critical loop has
the highest priority, the user interface thread that updates front panel objects
cannot use any processor time. As a result, the front panel and VI appear locked.
However, the code is still running.

While msec to wait control is greater than 0, the time-critical loop will “sleep”
for that many milliseconds. During this sleep, you can run non time-critical tasks
like updating user interfaces.

9. Click the Abort button. After a short period of time, the RT Development System
recognizes that it has lost communication with the RT Series device, and displays
a warning message. Click OK to close the RT Development System. You must
reset the RT target before the RT Development System can communicate with the
target platform again.
10. Exit LabVIEW Real-Time. Do not save any changes.

Exercise 3-2
Objective: To build a Multi-threaded RT Engine VI that communicates with
the Host PC through VI Server.

1. Build the following VI and set it to Time Critical Priority. Save the global
variables in a file called Global.vi. This Time Critical Loop represents a VI
that needs to run in real-time. It generates “data” in the form of a numeric.

 National Instruments Corporation 3-23 LabVIEW Real-Time

 We would like to monitor the data from the Host PC, so we will use global
variables for inter-thread communication to another VI that will assist in
communication with the Host PC. Remember to change the continuation terminal
of the while loop to Stop If True.

2. Next create a VI as shown below using the global variables from Global.vi.
Modify the connector for this VI so that stop is an input, and Numeric is an
output. This VI will be called from the Host PC using a Call By Reference
Node VI.

3. Build the following Top Level VI using Parameters VI and Time Critical Loop VI
as subVIs. This way you only have to download one VI instead of two.
Download and run your top level VI on your RT Engine.

LabVIEW Real-Time 3-24  National Instruments Corporation

 4. Check to make sure that VI Server on the RT Engine is configured to allow your
Host PC to access it. For RT Series DAQ Devices, this configuration is done
automatically. For RT Engines on the network, you will need to do the following:
a. With LabVIEW Real-Time targeted to the RT Engine, from a VI select
Tools»Network: x.x.x.x Options…
b. From the top pull-down menu of the Options dialog box, select
VI Server: Configuration and make sure that TCP/IP is checked
and that the Port is 3363.
c. From the top pull-down menu of the Options dialog box, select
VI Server: TCP/IP Access and make sure your Host PC’s IP address
is in the TCP/IP Access List. (Hint: if you put * in the access list,
then all clients are allowed VI Server Access.)
d. From the top pull-down menu of the Options dialog box, select
VI Server: Exported VIs and add * to the Exported VIs list.
e. Click OK when finished.

5. Switch execution target to the Host PC. Create the following VI. To create the
type specifier VI Refnum, right-click on that output terminal of the Open VI
Reference and select Create»Control. Right-click on the type specifier VI
Refnum (for type only) control and choose Select VI Server Class»Browse…,
then select the Parameters.vi. The type specifier VI Refnum tells
the Call by Reference Node what type of inputs and outputs are used by
Parameters.vi. Since the Parameters.vi VI is in memory on the RT Engine,
the Open VI Reference only requires the VI’s name, and not a path.

This VI will call Parameters VI remotely using a Call By Reference Node.

This will have the effect of writing the Stop Remote VI value to the stop global
variable on the RT Engine, and reading Numeric from the RT Engine.

 National Instruments Corporation 3-25 LabVIEW Real-Time

 6. Enter the IP Address or machine name for the RT Engine, and run the VI with
LabVIEW targeted to the Host PC.
7. Start and stop the Host PC VI and notice the behavior. The time-critical VI
is updating the Numeric global variable on the RT Engine. When we call
Parameters VI, it reads that global variable on the RT Engine and returns
it to the Host PC.
8. Click on Stop Remote VI, and then exit LabVIEW. This method
of communication is easy to program, but is used primarily for monitoring
the data on the RT Engine.

Exercise 3-3 (RT Engines on Network Only)

Objective: To build a Multi-threaded RT Engine VI that communicates
with the Host PC through a DataSocket Server.

1. Start the DataSocket Server on your Host PC by selecting Start»Programs»

DataSocket»DataSocket Server. If this option is not available, run
C:\Program Files\National Instruments\
DataSocket\cwdss.exe.
2. Start the DataSocket Server Manager on your Host PC by selecting Start»
Programs»DataSocket»DataSocket Server Manager. If this option is not
available, run C:\Program Files\National Instruments\
DataSocket\cwdssmgr.exe.
3. In the DataSocket Server Manager under Permission Groups, add “everyhost”
to DefaultReaders, DefaultWriters, and Creators, then select Settings»Save
Settings Now.
4. Build the following VI and use the Time Critical Loop VI and Global
VI that you created in Step 1 of Exercise 3-2.

LabVIEW Real-Time 3-26  National Instruments Corporation

 5. Enter dstp:\\[IP address]\Numeric (use your Host PC’s IP address)
for the URL of the DataSocket Item to write to.
6. Enter dstp:\\[IP address]\stop (use your Host PC’s IP address)
for the URL of the DataSocket Item to read from.
7. Run the VI targeted to your Networked RT Engine.
8. Switch execution target to the Host PC and build the following VI below.
9. Enter dstp:\\[IP address]\stop (use your Host PC’s IP address)
for the URL of the DataSocket Item to write to.
10. Enter dstp:\\[IP address]\Numeric (use your Host PC’s IP address)
for the URL of the DataSocket Item to read from.
11. Run the VI targeted to your Host PC.

 National Instruments Corporation 3-27 LabVIEW Real-Time

 12. Stop and start the Host PC VI and observe the behavior.
13. Click on the Stop Local VI button. This will stop the RT Engine VI.
14. Exit LabVIEW. Keep in mind when using DataSocket that the writer can write
more than once before a read occurs, resulting in lost data.

Exercise 3-4
Objective: To build a Multi-threaded RT Engine VI that communicates with
the Host PC through a VI Server and Get/Set Control Value.

1. Build the VI shown below. It is similar to the Time Critical VI used earlier. Set it
to time-critical priority.

LabVIEW Real-Time 3-28  National Instruments Corporation

 2. Check to make sure that VI Server on the RT Engine is configured to allow your
Host PC to access it. For RT Series DAQ Devices, this configuration is done
automatically. For RT Engines on the network, you will need to do the following:
a. First target LabVIEW Real-Time to the RT Engine, and from a new VI
select Tools»Network: x.x.x.x Options…
b. From the top pull-down menu of the Options dialog box, select
VI Server: Configuration and make sure that TCP/IP is checked
and that the Port is 3363.
c. From the top pull-down menu of the Options dialog box, select
VI Server: TCP/IP Access and make sure your Host PC’s IP address
is in the TCP/IP Access List. (Hint: if you put ‘*’ in the access list,
then all clients are allowed VI Server Access.)
d. From the top pull-down menu of the Options dialog box, select
VI Server: Exported VIs and add ‘*’ to the Exported VIs list.
3. Run the VI targeted to the RT Engine.
4. Target LabVIEW to the Host PC and build the VI below. The Flatten

To String VI, , and Unflatten From String VI, , are found in the
Functions»Advanced»Data Manipulation palette.

5. Enter the machine name or IP address of your RT Engine.

 National Instruments Corporation 3-29 LabVIEW Real-Time

 6. Run the VI in LabVIEW targeted to the Host PC.

When the Time Critical VI on the RT Engine is sleeping, the VI Server on the RT
Engine can communicate with the Host PC.

7. Stop and start the VI and observe the behavior.

8. Click on the Stop Remote VI button.
9. Exit LabVIEW.

LabVIEW Real-Time 3-30  National Instruments Corporation

 Lesson 4

Timing and Sleep

You Will Learn:

A. About the difference between software
and hardware timing
B. How to use various tools to control and
measure loop rates and induce sleep
C. About caching effects and how
to counteract them

ni.com

© National Instruments Corporation 4-1 LabVIEW Real-Time

 Software Determinism and Jitter

• Software jitter, though unpredictable,

can be measured, controlled, and bounded
• Use Wait Until Next ms Multiple to introduce
software timing and achieve µs jitter (~15 µs)

ni.com

We will soon see that we can easily implement software-timed control loops, where all
analog acquisitions and analog outputs are initiated through software. As a result we
are limited to the deterministic capabilities of software. Fortunately, with LabVIEW
Real-Time, software determinism is frequently adequate. By timing our loops with
Wait Until Next ms Multiple, we can mask out most of the software jitter within the
loop, leaving only the inherent software jitter incurred from the Wait function itself,
which is typically around 15 µs. One caveat about this method is that you must ensure
that your code within the loop executes faster than the requested Wait Until Next ms
Multiple period. This will become more clear once we have covered the Wait Until
Next ms Multiple function in detail.

LabVIEW Real-Time 4-2 © National Instruments Corporation

 Hardware Determinism and Jitter

• DAQ-STC has a 20 MHz time base

• Hardware jitter is typically on the order
of nanoseconds
• Current NI-DAQ versions* offer new event
messaging techniques using E-Series devices

ni.com

In systems that require more determinism (i.e., less jitter), we can use DAQ event
messaging techniques to achieve control response with nanosecond determinism.
By anchoring our timing to the DAQ time base, we can expect output responses
to occur at some user-defined interval, plus or minus a few nanoseconds!
By leveraging the determinism of the hardware, software jitter only becomes
a limiting factor of the overall performance of the control loop (i.e., it limits
the maximum control loop rate).
*NI-DAQ 6.7 introduced the AI SingleScan event message, and NI-DAQ 6.9
introduced the Counter Control—Wait event message.

© National Instruments Corporation 4-3 LabVIEW Real-Time

 Software vs. Hardware
Software Determinism
• Accurate loop timing achieved with Wait function
• Wait function masks software jitter within loop
but introduces a minor amount of its own jitter

Hardware Determinism
• Oscillator is independent from OS timing
• Hardware jitter depends on oscillator’s accuracy

ni.com

Before we move into great detail about timing and sleep, we can summarize the difference
between software and hardware determinism with a few key points. Software determinism
is typically much worse than hardware determinism because code execution time almost
always varies more than a timebase governed by a hardware oscillator (i.e., the 20 MHz
time base on the DAQ-STC). Depending on the time base and accuracy of the oscillator,
hardware jitter generally ranges from 1 to a few hundred nanoseconds. On the other hand,
software jitter within the real-time OS is around 15 µs at best. We will learn how to control
and minimize software jitter in this lesson and in Lesson 5. Once we have the tools to
measure software determinism, we can demonstrate these points more clearly.

LabVIEW Real-Time 4-4 © National Instruments Corporation

 Wait Until Next ms Multiple

1 ms OS updates

100 ms multiples

First wait Loop waits

(t)
100*N … … 100*(N+1)

First loop My Proc Second loop

cycle begins finishes and cycle begins
Wait begins

ni.com

Every millisecond, the OS updates an internal clock to keep track of time. Wait Until Next ms
Multiple synchronizes itself with this clock. When called, Wait Until Next ms Multiple waits
until the OS clock counts up to the next millisecond multiple requested. Each time Wait Until
Next ms Multiple is called in the above diagram, it waits until the OS clock has reached the next
multiple of 100 ms. If the loop begins executing arbitrarily between 100 ms multiples, Wait Until
Next ms Multiple waits long enough to synchronize the next loop cycle to the next 100 ms
multiple of the OS clock. In the slide shown above, My Proc may take a variable amount of time
to finish executing, but by calling Wait Until Next ms Multiple afterwards, we enforce a loop
frequency of 10 Hz (1/100 ms). The maximum loop rate that can be achieved using this method
is 1 kHz with a wait multiple of 1 ms. Since Wait Until Next ms Multiple only accepts integers,
loop rates are limited to only a few frequencies: 1000, 500, ~333, 250, 200, ~167 Hz, etc.
In the example above, the 100 ms timer is initialized by calling the Wait function immediately
before the while loop begins. Otherwise, the loop time for the first iteration would be
indeterminate. In the while loop, the Wait function is enclosed in a sequence structure to
assure that the delay is added after My Proc has finished. The output of My Proc is wired to the
sequence structure to create data dependency. In this way, the order of execution is guaranteed.
Before deciding on a millisecond multiple, you must be certain that the code in your loop can
execute faster than the wait multiple. If your code takes slightly longer than the ms multiple, then
your loop will actually wait two ms multiples, since code was running when the first ms multiple
arrived. In this case, the wait until ms multiple will not be aware that the first ms multiple
occurred, and will wait until a second ms multiple occurs before returning.
In addition to controlling loop rates, Wait Until Next ms Multiple forces time-critical VIs to
“sleep” until the wait is finished. When a VI sleeps, it relinquishes the CPU, allowing other VIs
or threads to execute. Sleep time is required in time-critical VIs because the user interface and
other background processes (ftp server, etc.) need CPU time to survive. Without any sleep,
a time-critical VI starves other processes, including the user interface and communication thread
in development mode.

© National Instruments Corporation 4-5 LabVIEW Real-Time

 Software-Timed Analog Input

Accurate software timing

requires Twc be less than
the wait multiple, ∆T.

(t)
(SW timing, ∆T = 1 ms ± inherent jitter)

(Code execution time, Te)

(Worst case software jitter, Tj)

Wait (Worst case time, Twc + wait)

& sleep
ni.com

The Wait Until Next ms Multiple function masks software jitter within the loop. The Wait
function has some inherent jitter, which is acceptable for many real-time applications.
In this slide, the Wait function synchronizes with each 1 ms tick of the OS clock, allowing
the loop to achieve 1 kHz software-timed analog input. The timeline illustrates the 1 ms
wait multiple in yellow (top arrow), the actual time required to execute the code in red,
worst case jitter in purple, and worst case time in green. As long as the worst case time
is smaller than the wait multiple, the actual loop period will equal the wait multiple,
plus or minus the jitter incurred by the wait function itself.

LabVIEW Real-Time 4-6 © National Instruments Corporation

 AI SingleScan Event Messaging

Scan clock wakes up

AI SingleScan

AI SingleScan sleeps and

waits for next scan

(t)
Wait AI SingleScan Wait AI SingleScan
executes executes

First loop cycle First loop cycle

begins finishes

ni.com

Using the Wait Until Next ms Multiple function, you can achieve a loop rate of up to 1000 Hz.
Alternately, loop rates can achieve their maximum rate if no delay is present, but without
a delay, you cannot precisely control the rate of the loop. Furthermore, if the VI is set for
time-critical priority, you risk starving other processes of CPU time. NI-DAQ 6.7 introduced
AI SingleScan event messaging, which is almost identical to the existing DAQ occurrences.
AI SingleScan event messaging, however, is faster and easier to program than DAQ occurrences
because no extra VIs are required to anchor loop timing to the scan clock signal on E-Series
DAQ devices.
As you may have already noticed, the loop rate matches the scan rate programmed at AI Start,
and this means your loop rates are no longer restricted to 1000 Hz. The scan clock is simply
an onboard signal which initiates analog to digital (A/D) conversions at the user-defined scan
rate. Each scan clock edge corresponds to a single scan across all channels in the scan list.
Calling AI SingleScan can have many different effects, depending on how the buffer and
hardware have been configured. To activate AI SingleScan event messaging, you must configure
the scan clock (i.e., set up hardware timing as opposed to software timing), and you must use
no buffer (i.e., buffer size = 0). In “hardware-timed, non-buffered” mode, AI SingleScan collects
data from the onboard FIFO buffer each time it is woken up by a new scan clock edge. For
relatively slow scan rates, AI SingleScan can retrieve new data as fast as it enters the onboard
FIFO. Overwrite errors can occur, however, when the requested scan rate is too fast relative to
the code execution time. In other words, if the CPU is not powerful enough to execute the code
within the loop at least as fast as the scan rate, the FIFO will eventually overflow with new,
unread data.

© National Instruments Corporation 4-7 LabVIEW Real-Time

 Hardware-Timed Analog Input

Accurate hardware timing

requires Twc be less than
the scan interval, ∆T.

(t)
(HW timing, ∆T = scan interval ± jitter)

(Code execution time, Te)

(Worst case software jitter, Tj)

Wait (Worst case time, Twc + wait)

& sleep
ni.com

AI SingleScan acts like Wait Until Next ms Multiple because both functions synchronize themselves
to a clock, and both functions induce sleep within the calling VI during the wait period. Unlike Wait
Until Next ms Multiple, however, AI SingleScan has an output parameter, which indicates when the
loop is keeping up with the requested timing interval. “Data remaining” is a convenient indicator of
whether the overhead in the loop is preventing AI SingleScan from keeping up with the scan clock.
You can verify that AI SingleScan’s sleep mode is enabled by looking for the following entry in your
NI-RT.INI file associated with the target on which you are executing code:
[NI-DAQ]
UseSleepingTimedNonBuffered=TRUE
This entry is present by default for all networked RT Series PXI Controller targets. Networked RT
Series PXI Controller’s NI-RT.INI files can be found in C:\NI-RT.INI of the target’s hard
drive. The AI Single Scan sleep mode is not turned on by default for the 7030 family RT Series DAQ
Devices. The 7030 product family’s NI-RT.INI file can be found in your ~\LabVIEW
RT\Resource folder.
In addition to providing greater-than-millisecond resolution to control loop timing, AI SingleScan's
“sleep mode” benefits multithreaded applications. A control loop running at time-critical priority
will starve other low priority threads if there is no delay or sleep mode built into the loop.
AI SingleScan’s sleep mode ensures that your time-critical code relinquishes the CPU while it is
waiting for the next scan clock edge, allowing other lower priority processes, including the user
interface, TCP/IP communication, and file I/O to run. Furthermore, only the time-critical priority
setting ensures a 1-to-1 correlation between your control loop and scan clock signal. If set to normal
priority, for instance, AI SingleScan is allowed to read from the onboard FIFO when a scan clock
edge arrives OR when there is a backlog of data in the FIFO. This is important because the “scans
remaining” output parameter of AI SingleScan can be misleading when run at normal priority.
Because AI SingleScan is not necessarily 1-to-1 with the scan clock, there is no guarantee that
zero scans remaining means AI SingleScan never missed a scan clock edge. Note: When using
hardware-timed, non-buffered AI SingleScan, you must be sure that no VIs are called between AI
Start and the first call to AI SingleScan to avoid a backlog of scans in the FIFO. Later, we will learn
about the input “read newest data,” which can counteract this effect in certain cases.
LabVIEW Real-Time 4-8 © National Instruments Corporation
 Measuring Loop Rates—Software
Subtracting two well-placed Tick Counts can yield
accurate timing information.

ni.com

The easiest method for measuring how fast a loop executes involves Tick Count.
By simply taking a tick count before and after several iterations of a loop, we can
estimate average loop cycle time with millisecond resolution. In this slide, we use
a sequence structure to ensure Tick Count gets called immediately before and
immediately after the loop we are timing.

© National Instruments Corporation 4-9 LabVIEW Real-Time

 Exercise 4.1
Students control loop cycle time with AI Single Scan
event messaging, and then measure average cycle
time with Tick Count.

Time to complete: 30 minutes

ni.com

LabVIEW Real-Time 4-10 © National Instruments Corporation

 Hardware-Timed Analog I/O

System under control experiences only hardware jitter

when AI and AO share a clock.

ni.com

By leveraging the determinism of the hardware, software jitter only becomes

a limiting factor of the overall performance of a control loop, limiting the maximum
control loop rate, not the determinism felt by the system under control. Software
jitter is no longer an issue for the system under control because the system under
control only experiences hardware jitter.
This slide depicts a modified example, which uses AI SingleScan event messaging.
At very accurately spaced hardware clock cycles, analog inputs and outputs are
acquired/updated simultaneously because they are sharing the same clock on PFI7.
The diagram ensures that updates sent to the system under control occur exactly
when scans occur with only a few nanoseconds of jitter. Therefore the system being
controlled by LabVIEW Real-Time will not experience any software jitters long as
the nominal “Code Execution Time,” Te, plus the “Worst Case Jitter” spike, Tj,
is less than the “HW scan interval,” ∆T. If this condition is always true, the system
under control will always experience hardware determinism, which is on the order of
nanoseconds!

© National Instruments Corporation 4-11 LabVIEW Real-Time

 Hardware-Timed Analog I/O
Scan clock initiates A/D
and D/A and wakes up AI
SingleScan concurrently

AI SingleScan sleeps and

waits for next scan

(t)

Wait AI SingleScan collects Wait AI SingleScan collects point

point 1 from FIFO, then 2 from FIFO, then AO
AO SingleUpdate loads SingleUpdate loads
response 1 for next D/A response 2 for next D/A

Loop cycle N Loop cycle N+1

begins begins
ni.com

This diagram illustrates the key ingredients to achieving hardware determinism with
AI SingleScan event messaging. As described on the previous slide, this diagram
features analog input and analog output sharing the same clock. AI SingleScan does
not initiate the A/D conversion, rather the scan clock does that. The same is true for
D/A conversions, which are initiated, not by the call to AO SingleUpdate, but rather
by scan clock. The system under control, therefore, is scanned and updated at
precisely the same time with each scan clock edge. Inside the loop, AI SingleScan
gathers a point (say point N) from the FIFO, which stores A/D samples. After the scan
is collected and processed, AO SingleUpdate loads a response to point N on the D/A
register. However, the D/A conversion of response to point N does not occur until the
next scan clock edge, when point N+1 is sampled. Using this architecture, D/A
response is delayed by one scan interval but is guaranteed to occur with only
a few nanoseconds of jitter.
Note: AO SingleUpdate must use the opcode “output only” when analog output
shares the scan clock with analog input.

LabVIEW Real-Time 4-12 © National Instruments Corporation

 Counter Control Event Messaging
Counter pulse wakes
up Counter Control

Counter Control sleeps

and waits for next pulse

(t)
Wait Counter Control Wait Counter Control
executes executes

First loop cycle First loop cycle

begins finishes

ni.com

As of NI-DAQ 6.9, Counter Control has a new parameter called “wait,” and Counter Set
Attribute has a new parameter called “timeout” (seconds). When performing pulse train
generation, Counter Control will induce sleep with the wait parameter and wake up when
the counter outputs the next pulse. To be precise, the VI wakes up one active source edge
(which we will call the waking edge) before the output begins the next pulse. If a waking
edge occurs before the Counter Control is entered, then the VI will return immediately.
So if waking edges are occurring too fast, your VI will not sleep at all. Because Counter
Control induces sleep, it can be a useful tool for preventing synchronous calls from hanging
your application.
Like AI SingleScan event messaging, Counter Control event messaging allows us to
anchor loop timing to hardware events. The only difference is Counter Control waits
for pulses generated by a general purpose counter on the DAQ-STC instead of pulses
generated by the scan clock, which wakes up AI SingleScan.

© National Instruments Corporation 4-13 LabVIEW Real-Time

 Exercise 4.2
Students control loop cycle time with Counter Control
event messaging, and then use Tick Count to measure
average cycle time.

Time to complete: 15 minutes

ni.com

LabVIEW Real-Time 4-14 © National Instruments Corporation

 Measuring Loop Rates—Hardware
Two ways to measure loop rates with Hardware:

1. Counter on DAQ board

2. Counter on CPU

ni.com

In some of the previous exercises, we used the Tick Count VI to time our loops. Though
Tick Count offers a simple way to measure average loop speed, it only provides us with
millisecond resolution. Consequently, calling Tick Count each time within the loop to
obtain an execution profile over time would not benefit us much if the loop were running
faster than 500 Hz. On the other hand, onboard general purpose DAQ counters and even
the CPU time-stamp counter offer much better timing resolution than 1 millisecond.
In fact DAQ-STC counters (i.e., those found on E-series devices) have up to a 20 MHz
time base, and NI-TIO counters, which are present on 660x devices, have a time base
of up to 80 MHz. If the RT target is a PXI controller, you can even use the processor’s
built-in time-stamp counter to keep track of time. The CPU counter increments with
each CPU clock cycle and has a resolution of 1/CPU frequency, which amounts to 3 ns
(1/333 MHz) on the PXI-8156B and ~1.17 ns (1/850 MHz) on the PXI-8170. In the next
few slides we will examine one way to time loops using counters and how to time loops
using the CPU’s time-stamp counter.

© National Instruments Corporation 4-15 LabVIEW Real-Time

 Measure Timing with a Counter
Configure a counter for simple event counting then
obtain a count within each iteration.

ni.com

By configuring a counter for simple event counting, we can utilize the accurate
time base of the DAQ-STC as the counter’s source. Once the counter has begun
incrementing, we can read the count register each time our loop iterates and store
these count values as time-stamps within an array for later processing. Subtracting
adjacent time-stamps yields approximate loop execution time for each loop iteration.
This information is more useful than an average loop cycle time because we can also
determine how much software jitter is present by plotting the time-stamps on a
histogram.
Of course, you have probably realized that by calling Counter Get Attribute within
the loop, we are adding overhead (and possibly jitter) to the actual loop execution
time. As an alternative to calling Counter Get Attribute, you can instead call Port
Write to toggle a digital line on the E-Series board. By wiring the output of the
digital line to the gate of a counter performing buffered event counting, you can
acquire timestamps with Counter Read Buffer.
Note: Buffered event counting is not covered in this course because it involves
advanced DAQ knowledge that is beyond the scope of this course. Please ask the
instructor to explain this method in more detail if Counter Get Attribute incurs too
much overhead relative to your actual loop execution time.

LabVIEW Real-Time 4-16 © National Instruments Corporation

 Measure Timing with RDTSC

Works on PXI systems but NOT 7030s!

ni.com

A faster, more precise timing method involves the CPU’s time-stamp counter. On PXI controllers
only, LabVIEW Real-Time users can call a special assembly instruction to measure loop timing at
the resolution of the processor clock (1/processor frequency). The assembly instruction is exposed
through a VI called Read Time Stamp Counter or RDTSC. This VI reads the 64-bit time-stamp
counter value from Pentium-class processors. The processor counter starts from zero each time the
controller is turned on. If the reset button is used to reset the controller, the processor counter will
reset. If MAX is used to reset or reboot the controller, the processor counter will reset. In order to
convert the time stamp information to a time value, you need to input the processor speed of your
machine to this VI. The timing code takes the array of time-stamps and converts them into actual
loop cycle times based on the CPU speed. The results are then plotted as Loop Cycle Time vs.
Loop Iteration.
This VI can yield erroneous data when run on laptops because laptops frequently adjust the
processor clock speed dynamically to conserve energy when there is no load. To remedy this
problem, place an empty running WHILE loop set at NORMAL PRIORITY in parallel with the
application being timed. The RDTSC instruction also returns erroneous data from dual processor
machines because the instruction does not always grab the time stamp from the same processor.
Each processor's Time Stamp Counter runs independently of the other.
Note: AMD processors used in the PXI-8156B Controller are Pentium-class.

© National Instruments Corporation 4-17 LabVIEW Real-Time

 RDTSC—Front Panel Results

ni.com

The timing shell’s front panel consists of several controls and indicators, most
notably the one large graph. The most relevant controls include iterations and
processor speed. Iterations dictates the number of test code iterations the FOR loop
will execute, and processor speed lets the timing code accurately convert time-stamp
data into actual loop cycle times. As you can see, the timing shell displays loop
cycle time for each loop iteration.

LabVIEW Real-Time 4-18 © National Instruments Corporation

 Exercise 4.3
Students examine RDTSC benchmark utility and time
various diagrams.

Time to complete: 40 minutes

ni.com

© National Instruments Corporation 4-19 LabVIEW Real-Time

 Memory Cache
Cache memory lets CPU retrieve
instructions and data quickly.

The CPU may swap new instructions

between main memory and cache.

Old Instructions

Main Cache
Memory Memory

New Instructions
ni.com

The most commonly used instructions are kept in a small cache, generally made up
of SRAM located on or near the processor die. Level 1 cache typically resides on the
processor die, while level 2 cache is separate from the processor but can be accessed
faster than normal memory. When new instructions must be processed, caching could
occur and might affect execution times. Processor are designed to optimize execution
time by placing more commonly used instructions in cache memory, which can be
accessed faster than main memory.

LabVIEW Real-Time 4-20 © National Instruments Corporation

 Caching Effects
The first few cycles of your control code might run slowly.

Read Newest
Data
Read Oldest
Data

Use AI SingleScan’s opcode to counteract caching effects.

ni.com

Processor cache effects the speed that code can execute in a loop. The first time the loop is
run, the cache does not necessarily contain the instructions that are used in the loop. Instead
the processor most likely still has instructions that were most frequently used before the loop
began running. When instructions that the processor needs are not in the cache, it takes longer
for them to load, but if they are in the cache, they load quickly from the cache.
This means that your initial loops may execute slower than later loops.
If you are using software or hardware timing to time your loops, and the loop execution time
is much faster than wait multiple (software timing) or scan interval (hardware timing), then
you do not need to worry about caching effects. However, in order to achieve optimal loop
rates, you will need to account for caching effects.
With AI SingleScan, you can counteract caching affects by using the opcode. For the first few
iterations, use the opcode “read newest data.” This assumes the following:
• Buffer Size = 0 (Sleeping will not work otherwise.)
• Hardware Timing
• RTOS
• Sleep Mode is enabled
• AI SingleScan is in a loop in a Time Critical VI
Read Newest will return the newest scan in the FIFO when the scan clock rising edge
occurs and will empty the FIFO. After each loop iteration, the loop will execute faster and
asymptotically approach some maximum loop rate. When the loop rate is running faster than
1/scan rate, you can then switch the opcode to read oldest data.

© National Instruments Corporation 4-21 LabVIEW Real-Time

 Counteracting Caching Effects
Read newest data
to collect most recent
scan from the FIFO.

Remaining stale scans

are thrown out.

Stale scans
(t) remaining in FIFO
…
Newest scan
collected from FIFO
AI SingleScan collects newest
scan from FIFO and removes
stale scans FIFO

ni.com

Caching effects generally disappear after the first few iterations unless you have other
threads running while your time-critical code sleeps. Consider a scenario where you
have achieved an optimal loop execution time in a time-critical loop that sleeps a lot.
You also counteracted initial caching effects by switching from read newest to read
oldest after many iterations. Now if you introduce lower priority code that can run
while your time-critical code is sleeping, it will invalidate your cache, causing the
processor to swap instructions back and forth between main memory and cache
memory each time LabVIEW RT switches threads. However, since the time-critical
code is already running at near optimal speeds, it will have less time to sleep when
caching occurs. As the time-critical code sleeps less, the other lower priority threads
are allotted less CPU time or possibly none at all. In the worst case scenario, the time-
critical loop will not keep up with the wait multiple or scan interval, depending on
whether you have implemented software or hardware timing. When the loop execution
time cannot keep up with the requested loop interval due to caching effects, you must
slow down the loop timing. Generally, you do not have to compromise your loop rate
to run lower priority threads. This is only true if caching effects prevent your time-
critical code from keeping real-time.
Read newest data with AI SingleScan throughout the initial iterations. Using the
“read newest data” opcode, only the newest scan in the FIFO is collected from the
DAQ device, while the remaining points are thrown out. Reading newest data prevents
the FIFO from overflowing during the caching iterations and removes stale data points,
which can disrupt control response.

LabVIEW Real-Time 4-22 © National Instruments Corporation

 Counteracting Caching Effects
After several iterations,
read oldest data.

Stale scans left in the

FIFO are called “scans
remaining.”

Stale scans
(t) remaining in FIFO
… …
Newest scan
collected from FIFO
AI SingleScan collects oldest
scan from FIFO and leaves
stale scans if present

ni.com

After initial caching effects have subsided, switch AI SingleScan’s opcode

to “read oldest data.” By switching the opcode, we can now monitor AI SingleScan’s
“scans remaining” output parameter. This parameter is a measure of real-time because
scans remaining must always be zero to ensure our loop execution time occurs faster
than the requested scan interval. On the other hand, if scans remaining grows larger
than zero, we can deduce that AI SingleScan is not keeping up with the scan interval
(i.e., not keeping real-time).

© National Instruments Corporation 4-23 LabVIEW Real-Time

 Counteracting Caching Effects

Use the iteration terminal to switch the opcode.

ni.com

LabVIEW Real-Time 4-24 © National Instruments Corporation

 Cache Effects and RDTSC
Use “Initial points to ignore” in RDTSC utility to
only analyze data after caching effects.

Discard Keep

“N” initial loop iterations are thrown out of the timing code
before mean cycle time and jitter are evaluated.

ni.com

The RDTSC timing shell has provisions for initial caching effects. The “Initial
points to ignore” allows you to discard the first “N” time-stamps so as not to group
the caching iterations with the optimal loop iterations when running the timing code
to compute mean, minimum, and maximum loop cycle time.

© National Instruments Corporation 4-25 LabVIEW Real-Time

 Chapter 4—Summary
!Software versus hardware jitter
• Software jitter is usually larger than hardware jitter
• Hardware timing masks software jitter
!Controlling and measuring loop rates
• Use Wait Until Next ms Multiple to control loop rate
• Use Tick Count to measure loop rate with ms resolution
• Use onboard counters to control or measure loop rate
• Use RDTSC to measure loop rate—highest resolution
!Caching Effects can induce jitter
• Use AI Single Scan “read newest data” to compensate
for caching
• Use “Read N Extra Initial Entries” to avoid caching
in RDTSC
ni.com

LabVIEW Real-Time 4-26 © National Instruments Corporation

 Exercise 4-1
Objective: To control loop rates by using AI Single Scan.

1. Launch LabVIEW Real-Time and select the appropriate RT Series Device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. For RT Series DAQ Devices, select the Reset checkbox to reboot the hardware.
3. Select File»Open and open Cont Acq&Chart (hw timed) VI from c:\Program
Files\National Instruments\LabVIEW 6\Examples\daq\
anlogin\anlogin.llb.
4. Right-click on the VI icon on the front panel and select VI Properties.
5. Select Execution Options from the pull-down menu. Set the priority to time
critical priority (highest). This makes the VI run in the highest priority thread
on the RT Engine.

To achieve real-time control, we must run the VI at time-critical priority to ensure

that 100% of the CPU time is dedicated to acquiring and processing within the
loop. AI Single Scan will synchronize itself to the scan rate and will also provide
sleep within the loop when the software finishes before the next scan clock edge.
This sleep time yields CPU time to other lower priority processes like the
communication thread that sends new values to our user interface running
on the host machine.

Before we run the example, we will add a Tick Count before and after the loop
to determine whether AI Single Scan really is synchronized with the scan clock.

6. Open the functions palette by right-clicking on the diagram window and select
the sequence structure located under the structures sub-palette.
7. Wrap the sequence structure around the while loop in the example.

 National Instruments Corporation 4-27 LabVIEW Real-Time

 8. Next add a sequence frame before and after the while loop and place Tick Counts
in these new sequence frames. In frame two, perform a subtraction on the Tick
Count values to determine approximate execution time.

This timing shell will only tell us how long overall the code took to execute.
To determine how fast each loop cycle executed on average, we need to divide
the overall time by the number of loop cycles. This can easily be accomplished by
passing the loop iteration terminal from the while loop in frame 1 to the timing
code in frame 2.

9. Modify the code in frame 1 and frame 2 to determine the average execution time
per loop cycle.
10. Run the VI with a scan rate of 100 Hz, and then stop the VI after a few seconds.
Notice that the average loop cycle time was approximately 1/100 second, which
corresponds to a period of 10 milliseconds.

Exercise 4-2
Objective: To control loop rates by using a DAQ-STC counter.

1. Launch LabVIEW Real-Time and select the appropriate RT Series Device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. For RT Series DAQ Devices, select the Reset checkbox to reboot the hardware.
3. Select File»Open and open counterwait VI from Exercises\
Chapter 4\counterwait.llb.

LabVIEW Real-Time 4-28  National Instruments Corporation

 4. Study the diagram and notice that the VI simply performs pulse train generation
with a couple of modifications.

The first thing that sets this example apart from normal pulse train generation
is that we configure a timeout (in seconds) using Counter Set Attribute VI.
The second node that makes this example unique is Counter Control within the
while loop. Notice that we use the wait parameter, which causes Counter Control
to become synchronized to the pulse train generated by the counter. If for some
reason the counter failed to start, we need it to time out to prevent the loop from
hanging indefinitely, hence the timeout mentioned earlier.

5. Now add a sequence timing shell around the while loop as you did with AI Single
Scan in the previous exercise.
6. Experiment with different pulse spec parameters and time the loop for several
iterations to verify that Counter Control really is synchronized to the counter’s
pulse train.

Exercise 4-3
Objective: To examine the Read Time Stamp Counter (RDTSC) timing
shell. (For RT Series PXI Controllers Only)

1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. Select File»Open and open RDTSC Timing Shell VI from
Exercises\Chapter 4\RDTSC Timing Shell.llb.
3. Look at the diagram. Inside the VI, you will find a For Loop and a subVI called
RDTSC.vi, which is merely a call to some external assembly-level code that reads
the time-stamp count register on Pentium processors. RDTSC returns two 32-bit
numbers corresponding to the upper and lower halves of a Pentium processor’s
64-bit time stamp counter register. Each time the Pentium clock generates a new
clock edge, the time stamp counter register increments once. By tracking the
counter register each time our loop executes, we can extract valuable timing
information with a resolution equal to 1/CPU speed. For our fastest controller,
which has an 850 MHz processor, we can measure loop rates with nanosecond
timing resolution.

When timing code with RDTSC, it is best to ensure that RDTSC executes at
exactly the same time with respect to the rest of the code inside your loop to
ensure accurate timing data is collected. To achieve this effect, we recommend
you wire RDTSC to your code such that RDTSC is either the last node of the first
node to execute within your loop.

 National Instruments Corporation 4-29 LabVIEW Real-Time

 The remaining parts of the diagram convert RDTSC data to actual times based
on the CPU speed. The timing data is then displayed on a graph with cursors
indicating the mean loop execution time and standard deviation.

4. Use RDTSC Timing Shell to time simple multiplication.

5. Time Cont Acq&Chart (hw timed) VI using the RDTSC Timing Shell.

LabVIEW Real-Time 4-30  National Instruments Corporation

 Lesson 5

Shared Resources

You Will Learn:

A. About different types of shared resources
B. About priority inversions and priority
inheritance
C. About communicating between VIs using
queues and functional global variables

ni.com

© National Instruments Corporation 5-1 LabVIEW Real-Time

 Shared Resources
A shared resource in LabVIEW RT is anything that
can only be used by one process at a time.

LabVIEW RT shared resources include:

– Global variables
– LabVIEW RT Memory Manager
– Non-reentrant subVIs
– Single-threaded DLLs
– Semaphores
– Networking code (TCP/IP, UDP, VI Server)
– File I/O

ni.com

Certain data structures, driver libraries, and variables can only be accessed serially,
one process at a time. A simple example of a shared resource common to all
programming languages is the global variable. As you know, global variables cannot
be accessed simultaneously from multiple processes to prevent race conditions.
Therefore, compilers automatically protect the global variable as a shared resource
while one process needs to access it. In the meantime, if a second process tries to
access the global variable while it is protected, the second process must wait until the
first process has finished with the global variable. As we will soon see, understanding
shared resources and how to identify them is an important skill when programming
real-time applications.
Note: These operations are inherently non-deterministic and should never be used
inside a time-critical priority loop if you’re attempting to achieve real-time
performance.

LabVIEW Real-Time 5-2 © National Instruments Corporation

 Shared Resources
Before a process can begin using a shared resource, it
must obtain a mutual exclusion or mutex.

Process 1 Shared Process 2

Resource
Running Waiting

After Process 1 finishes, Process 2 can proceed.

Process 1 Shared Process 2

Resource
Waiting Running

ni.com

For our purposes a shared resource is defined as a software object that can only be used
by one thread at a time. In this slide, let’s assume both processes constantly need to
access the shared resource. Let’s also assume Process 1 is running at normal priority,
while Process 2 runs at time-critical priority. Normally, when a time-critical thread
needs to execute, it preempts all other code running on the real-time target; however,
a normal priority thread can block the time-critical thread if it has not released a mutex
that the time-critical thread needs. This is known as a priority inversion because the
real-time operating system cannot allow the time-critical thread to preempt the normal
priority thread, merely because of the mutex around a shared resource.
A typical scenario of a priority inversion caused by a shared resource follows:
• When a lower priority thread needs a resource such as the LabVIEW Memory
Manager, it protects the resource from external access while in use. The lower priority
thread has acquired a mutex and begins allocating memory.
• Now, a time-critical priority thread wakes up while the lower priority thread
is using the resource and kicks the lower priority thread off of the processor.
• Then the time-critical priority thread wants to use the same protected resource
(i.e., wants to allocate memory), but is forced to wait because of the protection
(the mutex) around the resource. This is an inversion of priorities; a higher priority
thread is forced to wait because a lower priority thread has protected a resource the
higher priority thread needs.

© National Instruments Corporation 5-3 LabVIEW Real-Time

 Priorities and Shared Resources

ni.com

In this slide the shared resource is a global variable, which is shared by two VIs:
one set to normal priority and one set to time-critical priority.

LabVIEW Real-Time 5-4 © National Instruments Corporation

 Priority Inversion
Time Critical Priority - waiting
Priority inversion:
Normal priority VI blocks the
higher priority VI with a mutex
around the shared resource.

Normal Priority - running

Priority inheritance:
Normal priority VI inherits higher
priority to release mutex.

ni.com

The real-time OS uses a method called priority inheritance to resolve the priority inversion
as quickly as possible. It does this by:
• Allowing the lower priority thread to temporarily “inherit” the time-critical priority setting
long enough to finish using the shared resource and to remove the protection.
• Once the protection is removed the lower priority thread resumes its original lower priority
setting and is taken off of the processor.
• Now the time-critical priority thread is free to proceed and use the resource (i.e., access the
global variable).
A result of this priority inversion is increased software jitter in the time-critical priority thread.
The jitter induced by a protected global variable is generally quite small compared to the jitter
induced by protecting the LabVIEW Memory Manager. Unlike accessing global variables,
performing memory allocations is unbounded in time and can introduce a broad range of
software jitter while parallel operations try to allocate blocks of memory in a wide variety of
sizes. The larger the block of memory to be allocated—the longer the priority inheritance takes
to resolve the priority inversion.

© National Instruments Corporation 5-5 LabVIEW Real-Time

 Sharing a Global Variable
Only one loop at a time has access to the global variable
“stop.”

1) Top loop acquires a mutex to the

global and writes to it.

2) Lower loop becomes locked

while waiting to acquire mutex.

3) Top loop releases mutex around

global, allowing lower loop to
proceed.

ni.com

The slide animation depicts one loop acquiring a mutex to the global variable,
blocking the other loop for a brief moment. The global variable is a very common
type of shared resource, which can cause problems for real-time applications. Every
shared resource conflict inevitably manifests itself as software jitter. Therefore, we
must learn how to avoid priority inversions and shared resource conflicts in general
to minimize software jitter.

LabVIEW Real-Time 5-6 © National Instruments Corporation

 Functional Global Variables (FGV)
FGVs can have
several inputs and
outputs.

FGVs can be
“skipped if busy.”

ni.com

Instead of using regular global variables, LabVIEW Real-Time programmers should

employ functional global variables to minimize software jitter. A functional global variable
(or LabVIEW 2-style global) is merely a subVI set at subroutine priority that contains a WHILE
loop with a nested case structure for read or write access. FGVs always run once when they are
called and store data in their shift registers, which allow subsequent FGV calls to access the most
recent data.

© National Instruments Corporation 5-7 LabVIEW Real-Time

 Functional Global Variables
A time-critical thread
should be allowed to
skip the FGV call.

When skipped, the FGV

returns default values
for each data type.

ni.com

The “skip if busy” setting ensures that the time-critical thread skips the FGV if another process
is currently accessing the same FGV. When skipped, the FGV returns default values for each
data type on the FGV diagram: numerics return 0, Booleans return false, etc. The FGV images
used in the slides depict only one set of shift registers for “A” and “B.” FGVs are similar to
queues because you can add more shift registers to store a longer history of values of “A” and
“B.” You can also add more variables to the FGV, to store “C,” “D,” “E,” etc. An advantage
of FGVs over normal global variables is that FGVs help prevent shared resource conflicts while
still offering the same functionality as normal global variables. A disadvantage to FGVs is that
you may lose data during inter-thread communication when the FGV gets skipped inside the
time-critical thread.

LabVIEW Real-Time 5-8 © National Instruments Corporation

 Exercise 5.1
Students measure software jitter incurred from
global variables and learn how to use Functional
Global Variables (FGVs) for inter-thread
communication.

ni.com

© National Instruments Corporation 5-9 LabVIEW Real-Time

 Sharing SubVIs
You can configure subVIs for reentrancy to avoid shared
resource conflicts.

1) Top loop acquires a mutex to the

VI and writes to it.

2) Lower loop becomes locked

while waiting to acquire mutex.

3) Top loop releases mutex around

VI, allowing lower loop to
proceed.

ni.com

As you saw in the previous slides, sharing subVIs cause priority inversions the same way
global variables do. When a subVI is set to subroutine priority, that subVI can be “skipped”
within time-critical code to avoid software jitter that would have occurred from a priority
inversion. If you are running unrelated parallel processes that call the same subVI, you can
configure the subVI for reentrant execution. A reentrant subVI establishes a separate data
space in memory each time it is called. This feature allows LabVIEW RT to call several
instances of a particular subVI simultaneously. Because reentrant subVIs use their own data
space, however, they cannot be used to share or communicate data between threads. You
should only use reentrancy when you must simultaneously run two or more instances of a
subVI within unrelated processes that do not need to share data within the reentrant subVI.

LabVIEW Real-Time 5-10 © National Instruments Corporation

 Reentrant SubVIs
Configure reentrancy in the Execution window under VI
Properties.

Reentrant subVIs do NOT act like global variables!

ni.com

To configure a subVI for reentrancy, go to VI Properties»Execution, and then check

the box “Reentrant Execution.” Be careful when using this feature because each call to
the subVI establishes a new unique data space for all controls and indicators in memory.
Reentrant subVIs cannot be used as LabVIEW 2-style global variables for this reason.
You should also be aware that configuring the DAQ VIs for reentrancy does not solve
shared resource problem with NIDAQ32.dll. As we will soon see, all DAQ VIs load the
DAQ driver, NIDAQ32.dll. This dll is not multithreaded safe, which means it cannot be
called simultaneously from multiple processes.

© National Instruments Corporation 5-11 LabVIEW Real-Time

 Sharing the DAQ Driver
NI-DAQ32.dll is single-threaded,
making it a shared resource.

1) Top loop acquires a mutex

to the DLL.

2) Lower loop becomes locked

while waiting to acquire mutex.

3) Top loop releases mutex

around DLL, allowing lower
loop to proceed.

ni.com

As mentioned in the previous slide, NIDAQ32.dll is not multithread safe. This has
serious implications for parallel DAQ processes because the DAQ driver can only be
accessed by one process at a time, much like a global variable. Therefore, you must
architect your DAQ applications carefully to avoid parallel DAQ calls like the one
in this slide. The recommended method for architecting a DAQ application is to
call DAQ VIs in only one loop, serially. You can ensure that DAQ VIs are called
serially by routing one error cluster through every DAQ VI in the loop.

LabVIEW Real-Time 5-12 © National Instruments Corporation

 Data Acquisition
Key points to remember about DAQ and real-time:
• The NI-DAQ driver is a shared resource.
• All data acquisition VIs call NIDAQ32.dll.
• All DAQ calls should be made serially from within a
single loop within one thread.
• To achieve real-time, the DAQ loop must be given
time-critical priority.

ni.com

The only way to achieve true real-time DAQ, is to call all DAQ VIs from within the same
loop inside the same thread, which must be given time-critical priority.

© National Instruments Corporation 5-13 LabVIEW Real-Time

 LabVIEW RT Memory Manager
!LabVIEW RT manages memory automatically.
• The user does not explicitly have to allocate
or de-allocate memory.
• This means memory management is easy but harder
to control.
!LabVIEW RT has a “Memory Manager,”
which is a shared resource.
• We must control memory allocations to avoid shared
resource conflicts with the memory manager.
• Statically allocate memory before time-critical process
begins.

ni.com

LabVIEW RT has a memory manager, which automatically allocates and de-allocates

memory at run-time. Allocating memory can consume a significant amount of CPU time,
depending on the amount of memory needed. If we allow LabVIEW RT to dynamically
allocate memory at run-time, our application could suffer from software jitter for two reasons:
1. The memory manager may already be mutexed by another process, causing
a shared resource conflict.
2. Even if the memory manager is immediately available, allocating memory is
inherently non-deterministic because there is no upper bound on the memory
allocation’s execution time.
To avoid shared resource conflicts or, more specifically, priority inversions, with the
LabVIEW RT memory manager, statically allocate memory before any real-time loops.
For example, instead of using Build Array within your loop to index new data into an array,
use Replace Array Subset to insert new array values into pre-allocated arrays. We will learn
how to do this shortly.

LabVIEW Real-Time 5-14 © National Instruments Corporation

 Sharing the Memory Manager
Some LabVIEW RT functions allocate memory and must
therefore access the memory manager.

1) Top loop acquires a mutex to the Memory

Manager
memory manager.
Memory
2) Lower loop becomes locked
Manager
while waiting to acquire mutex.

3) Top loop releases mutex around

memory manager, allowing lower
loop to proceed.

ni.com

The slide features two parallel loops. The top loop dynamically builds an array
using the Build Array function, while the lower loop builds the same array using
auto-indexing.

© National Instruments Corporation 5-15 LabVIEW Real-Time

 Pre-allocate Arrays

Avoid shared resource conflicts

by pre-allocating arrays.

Initialize Array before entering the loop.

Replace Array Subset to insert new

elements.

ni.com

Initialize Array in combination with Replace Array Subset allows you to pre-allocate
arrays inside a loop to avoid shared resource conflicts with the LabVIEW RT memory
manager.

LabVIEW Real-Time 5-16 © National Instruments Corporation

 Memory Management
Key points to remember about memory
management and real-time:
• The LabVIEW RT memory manager is a shared
resource.
• All memory allocations must be removed from the
time-critical loop to guarantee robust, hard real-time
performance.
• Pre-allocate arrays outside of the time-critical loop.
• Cast all data to the proper data type.
• Use inplaceness when possible to reuse memory
buffers.

ni.com

In general, memory allocations within a time-critical loop will induce jitter and affect
the deterministic properties of a LabVIEW RT program. All memory allocations must be
removed to guarantee robust hard real-time performance. You must pre-allocate your arrays
outside of the loop if you want your application to run deterministically. Certain LabVIEW RT
functions, such as Build Array and Bundle, allocate memory.
Also try to cast data to the proper data type. Every time LabVIEW RT has to perform a type
conversion, either implicitly or explicitly, a copy of the data buffer is made in memory in order
to retain the new data type after the conversion. Use the smallest data type possible, and if you
need to convert the data type of an array, try to do that conversion before the array is built. Also
keep in mind that many LabVIEW RT functions reuse their input buffers when the output data
representation, size, and dimensionality matches the input. The same structure and number of
elements. This ability for a function to reuse buffers is called inplaceness.
Note: More information about inplaceness and LabVIEW memory management can
be found in the LabVIEW Advanced Course.

© National Instruments Corporation 5-17 LabVIEW Real-Time

 Exercise 5.2
Students learn how to pre-allocate arrays and
measure software jitter incurred from dynamic
memory allocations.

ni.com

LabVIEW Real-Time 5-18 © National Instruments Corporation

 Shared Resources and Handshaking
Avoid these functions inside time-critical code:
• Semaphores
• Networking functions (TCP/IP, UDP)
• File I/O
Avoid handshaking protocols too:
• GPIB
• RS-232 with h/w or s/w handshaking
• DAQ handshaking
• Any other handshaking protocol…

ni.com

We did not cover all types of shared resources in this chapter. Avoid semaphores, TCP/IP,
UDP, VI Server, and File I/O functions within your time-critical loop. These functions are
inherently non-deterministic and use shared resources. For example, semaphores are
themselves shared resources, network functions use the Ethernet driver, and file I/O functions
use the hard disk. These functions can introduce severe software jitter in time-critical code
on account of priority inversions.
Also be aware that all handshaking protocols are non-deterministic. GPIB, RS-232,
and TCP/IP are just a few things that should never be run at time-critical priority.
DAQ handshaking protocols like burst mode and 8255 emulation mode on the 653x boards
are also non-deterministic and should be avoided in time-critical loops.

© National Instruments Corporation 5-19 LabVIEW Real-Time

 Real-Time Queues (RTQ)
Real-Time Queues offer a safe method
of inter-thread communication.

ni.com

Real-Time Queues are like the regular LabVIEW Queues except they pre-allocate enough
memory required to store up to “n” elements of a user-defined numeric representation
(i.e., DBL, SGL, U32, etc.). As we learned earlier, pre-allocating memory is important because
dynamic memory allocations within a time-critical loop can add a significant amount of
software jitter. Each RT Queue allows users to push and pop elements simultaneously because
RT Queues actually consist of two queue copies in memory. Simultaneous read and write
access is a necessary feature of RT Queues to prevent shared resource conflicts. Like FGVs,
RT Queues allow you to safely pass data between threads. A benefit to using RT Queues instead
of FGVs is that you can pass data between threads without losing any data. Recall that FGVs
become “lossy” when they are skipped by the time-critical thread. One disadvantage to RT
Queues is that data cannot be accessed arbitrarily or in blocks. RT Queue elements must be
popped from the queue one element at a time.

LabVIEW Real-Time 5-20 © National Instruments Corporation

 Exercise 5.3
Students learn how to use RT Queues
for inter-thread communication.

ni.com

© National Instruments Corporation 5-21 LabVIEW Real-Time

 Chapter 5—Summary
A shared resource is a software object that
can only be used by one thread at a time.
• LabVIEW Memory Manager
• Global variables
• Non-reentrant subVIs
• Synchronization code
• Networking protocols
• File I/O

ni.com

LabVIEW Real-Time 5-22 © National Instruments Corporation

 Chapter 5—Summary (Continued)
!To avoid priority inversions:
• Pre-allocate arrays before entering loop
• Keep all DAQ calls in one time-critical thread
• Keep all networking and file I/O calls in a separate
normal priority thread
!To communicate data between threads:
• Use Functional Global Variables (FGVs)
• Or use RT Queues
!Avoid shared resources and handshaking
within time-critical code.

ni.com

© National Instruments Corporation 5-23 LabVIEW Real-Time

 Exercise 5-1
Objective: To transfer data between threads using a standard global
variables and functional global variables.

PART 1.
1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. Select File»New, and create a new VI set for normal priority that writes
to a global variable.

3. Select File»Open and open RDTSC Timing Shell VI from

Exercises\Chapter 5\RDTSC Timing Shell.llb.
4. Now open the RDTSC timing shell and modify it to write to the same
global variable. Add a small wait statement.

LabVIEW Real-Time 5-24  National Instruments Corporation

 5. Run the timing shell while globalWriter VI is running in the background.
6. Run the timing shell while globalWriter VI is idle and compare the timing
differences.

The timing shell should reveal slightly more jitter when globalWriter VI is
running in the background. If the global were much larger than a simple Boolean,
such as a large array or cluster, the jitter would be much more pronounced.

7. Try other types of global variables to see what effect they have on jitter.
Try keeping the user interface for the global VI open to see the effect on jitter.

PART 2.
1. Select File»Open and open Normal Priority Loop VI from
Exercises\Chapter 5.
2. Examine the block diagram and study LV2Global VI.

A Functional Global, also called a LabVIEW 2-Style Global, is a VI that uses

a loop with uninitialized shift registers to hold global data. A Functional Global
usually has an input parameter that specifies the VI function, such as writing
or reading. LV2Global VI is a Functional Global, with its priority set to
subroutine. When called, the subroutine receives 100% of the CPU time allocated
to its calling thread. While LV2Global VI is running, no other processes in the
same thread can execute. A benefit to using subroutine priority is that your code
can execute very quickly. Another benefit to marking a subVI for subroutine
priority is that you can skip over a call to the subroutine if it is currently being
used in another thread (i.e., if it is a shared resource). This option is called
“Skip Subroutine Call If Busy.”

 National Instruments Corporation 5-25 LabVIEW Real-Time

 3. Open Time Critical Loop VI. Copy the block diagram code into the
RDTSC Timing Shell so we can benchmark it. Make sure the “Skip Subroutine
Call If Busy” option is checked when you right-click on LV2Global VI.

4. Run Normal Priority Loop VI without modifications. Then run

the timing benchmark VI you’ve created.

LabVIEW Real-Time 5-26  National Instruments Corporation

 5. Now run the benchmark while the Normal Priority Loop is idle and compare
the timing differences. Generally, you’ll expect to see little difference in the jitter
when using Functional Globals because the time-critical VI will skip the call
if the subroutine is busy.

What are the timing differences between global variables and Functional Globals?
Which case had the least jitter?

Exercise 5-2
Objective: To examine the LabVIEW RT Memory Manager as a shared
resource using the RDTSC Timing Shell.

1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. Select File»Open and open RDTSC Timing Shell VI from
Exercises\Chapter 5\RDTSC Timing Shell.llb. Also open
Background Memory Allocation.vi.

PART 1.
3. Modify the RDTSC Timing Shell VI to build an array using the Build
Array function. This will cause the array to grow in size while the VI is running.
Use the Initialize Array function to start the array with a size of 10,000. Add
a small wait statement to the loop as well.

4. Run the Background Memory Allocation VI. Then run the timing
benchmark you’ve created.

 National Instruments Corporation 5-27 LabVIEW Real-Time

 5. Try running the timing benchmark with the Background Memory
Allocation VI stopped. Notice that the maximum jitter was greater when
the memory allocation VI was running.

You should avoid allocating memory within your time-critical code for two
reasons. First, other background memory allocation can cause a priority inversion,
which adds jitter to your loop. Second, the process of allocating memory is not
bounded in time. The time required to resolve a memory allocation depends on
the amount of memory requested and the amount of free memory available. As
you saw in this example, the amount of requested memory grows as we build
larger arrays while the amount of available memory diminishes.

LabVIEW Real-Time 5-28  National Instruments Corporation

 PART 2.
1. Modify RDTSC Timing Shell VI again to build an array using LabVIEW’s
auto-indexing feature instead of the build array function. Add a small wait
statement to the loop.

2. Run the Background Memory Allocation VI. Then run the timing
benchmark you’ve created.
3. Try running the timing benchmark with the Background Memory
Allocation VI stopped. Notice that the maximum jitter was not any greater
when the memory allocation VI was running. Memory is allocated before the
For Loop begins when using auto-indexing.

What conclusions can you draw from the timing results? How are memory
allocations handled differently with auto-indexing turned on than they are with
the Build Array function?

 National Instruments Corporation 5-29 LabVIEW Real-Time

 PART 3.
1. Modify RDTSC Timing Shell VI again to pre-allocate an array using
Initialize Array before entering the loop. Use Replace Array Subset to insert
random values into the initialized array. Include a small wait statement inside the
loop.

2. Run the Background Memory Allocation VI. Then run the timing
benchmark you’ve created.
3. Try running the timing benchmark with the Background Memory
Allocation VI stopped. Notice that the maximum jitter was not any greater
when the memory allocation VI was running. Memory is allocated before the
For Loop begins and the Replace Array Subset VI does not invoke the memory
manager.

How does performing memory management in a time-critical VI affect the

timing?

Exercise 5-3
Objective: To transfer data between threads using RT queues.

1. Select File»Open and open Simple RT Queue Example – Time Critical Loop VI
from Exercises\Chapter 5\RT queue\Simple RT Queue
Example - Time Critical Loop.llb.
2. Examine the block diagram and priority of this VI.

LabVIEW Real-Time 5-30  National Instruments Corporation

 This VI consists of two queues—Control Data Queue and Result Data Queue.
Simple RT Queue Example – Time Critical Loop VI reads two
elements from Control Data Queue and stuffs one element into Result Data
Queue.

3. Now select File»New, and build the following block diagram. Use the RT Queue
VIs from RTQ.llb.

The subVI node labeled “Real Time Loop” in the above illustration is merely
Simple RT Queue Example – Time Critical Loop VI. Simple
RT Queue Example VI pushes two new elements (a stop flag and amplitude)
into Control Data Queue and pops one element (sine data) from Result Data
Queue. When Result Data Queue has no new data available, the case structure
switches to TRUE and does nothing.

RT queues are superior to standard LabVIEW queues because they pre-allocate

memory based on queue size, number of queue elements, and the data
representation for each queue element. For example, Control Data Queue
defaults to a size of 10 elements and stores two DBLs per queue element.
Given these parameters, LabVIEW Real-Time pre-allocates 160 bytes

 National Instruments Corporation 5-31 LabVIEW Real-Time

 (8 bytes per DBL * 2 DBLs per element * size of 10 elements). By pre-allocating
queue memory, we avoid dynamic memory allocations, which are known
to induce jitter.

RT queues are also superior to normal global and functional global variables
because RT queues notify the user when they have lost data with the Boolean
parameter overwrite. When writing to global variables faster than reading them,
you will inevitably lose data, but there is no easy way to determine when exactly
this has happened.

RT queues can be complicated to new LabVIEW programmers. This drawback,

however, is generally outweighed by the advantages, which make RT queues
the preferred method for transferring data between threads within LabVIEW
Real-Time.

LabVIEW Real-Time 5-32  National Instruments Corporation

 Lesson 6

Other Features of LabVIEW Real-Time

You Will Learn:

A. How to automatically restart your application
if the RT target or host computer reboots
B. About the FTP server and Web server
running on RT PXI and FP-2000 targets
C. How to implement hardware and software
watchdogs

ni.com

© National Instruments Corporation 6-1 LabVIEW Real-Time

 Launching Application Builder
LabVIEW Real-Time’s Application
Builder compiles VIs into
executables.

ni.com

The LabVIEW Real-Time Application Builder compiles VIs into executables.

LabVIEW Real-Time executables can be automatically downloaded and run on
RT targets, which is an important feature for some users.

LabVIEW Real-Time 6-2 © National Instruments Corporation

 Application Builder Settings
The application
builder consists
of several tabs.

The tabs are used to target a VI for compilation, configure

exe settings, and even create an installer for the RT engine.
ni.com

We will not cover the details of Application Builder in this class. To learn more
about the Application Builder settings, please review LabVIEW Basics II.

© National Instruments Corporation 6-3 LabVIEW Real-Time

 App. Builder While Targeted
Application is downloaded to RT Device after
it is built.

ni.com

When targeted to a networked RT Series Device, after the application is built,

it is automatically downloaded to the Destination directory on the RT Engine.

LabVIEW Real-Time 6-4 © National Instruments Corporation

 Configuring RT Executables
When targeted to an RT PXI
Controller, the menu item “Network:
Options” under Tools becomes
enabled.

Check the “Launch Application at Boot-Up” box to

automatically launch exes when the RT target reboots.
ni.com

While targeted to an RT Series networked device, LabVIEW Real-Time offers special

menu items under Tools»Network: x.x.x.x Options. One of the network options is
RT Target: Miscellaneous. Selecting this option allows you to specify an application
path, which is the path our executable will be downloaded to. Also, if the Launch
Application at Boot-Up option is checked, the downloaded application will
automatically run each time the PXI device reboots.
To download an executable to an RT Engine immediately after building it, simply
invoke the Application Builder while targeted to an RT target. The executable’s
destination directory will change to the RT Engine’s network path instead of the local
machine. Upon rebooting the device, the application will automatically launch.
To turn off this feature, simply target LabVIEW Real-Time to the RT Engine, go to
Tools»Network Options»RT Target: Miscellaneous and deselect the check box
Launch Application at Boot-Up.

© National Instruments Corporation 6-5 LabVIEW Real-Time

 Command-Line Arguments

Use command-line arguments

within a shortcut to target an exe.
Use “-target,” “-quithost,” or “-reset.”
ni.com

Rather than downloading an executable to the RT-PXI hard disk, you can keep the
executable on the Host PC and run it targeted to the RT-PXI using a command line
arguments. This slide depicts the properties page of a shortcut to an executable called
Charts.exe. If we launch the shortcut, it will download and run Charts.exe on the
RT-PXI target at 130.164.48.131. By simply adding “-target” at the end of the
executable’s path, LabVIEW RT knows to download the executable to the RT-PXI target
and run it. The command line argument “-quithost” closes the user interface
to the executable on the Host PC and the communication thread on the RT-PXI.
To reset the device before or after running and executable, use “-reset.”
If the Host PC crashes while an executable is running on the RT Engine with a user
interface on the Host PC, if you later relaunch the executable targeted to the RT Engine,
the Front Panel will synchronize with the code still running on the RT Engine, and
continue to provide the user interface.

LabVIEW Real-Time 6-6 © National Instruments Corporation

 Command Line Arguments (Cont.)
Batch files and Windows
shortcuts can be placed
in the ~\startup directory
to automatically run a
targeted exe if the
Windows host reboots.

Batch files are an alternative to Windows

shortcuts. Batch files allow you to launch
more than one exe and still use command-line
arguments.
ni.com

Batch files offer an easy method of launching multiple executables to various targets
at once. You can create a batch file by simply opening a new text file with Notepad
and saving the document with the extension *.bat. Inside the batch file, type
command lines as though you were at a command prompt, supplying the full
path and name of the executable(s) you wish to launch, followed by any command
line arguments. Batch files and shortcuts are very handy when placed in the
~\Windows\startup folder, because they allow you to automatically download
and run executables to any target if the Host PC loses power.

© National Instruments Corporation 6-7 LabVIEW Real-Time

 Exercise 6.1

Students build an RT executable to

automatically start after power failure.

ni.com

LabVIEW Real-Time 6-8 © National Instruments Corporation

 FTP Server (PXI and FP-2000 only)

The PXI and FP-2000 Controllers run an FTP

server. Log on initially with no user name
or password.
ni.com

Logging on to the RT-PXI’s FTP server can be done from a command prompt or from
a Web browser window. In this slide, we are demonstrating how to log onto the FTP
server from the command prompt. To transfer files, you must have some knowledge
of FTP commands, which can be displayed by typing “?” at the FTP prompt. Be sure
to use DOS 8.3 file name format when specifying the path to the file you wish to
transfer. Example:
ftp> put c:\windows\progra~\nation~\labview\charts.vi
You should not exceed 8 characters per directory when specifying the path name.

© National Instruments Corporation 6-9 LabVIEW Real-Time

 FTP Server (Continued)

FTP transfer is easier with

a Web browser if you are not
familiar with FTP commands.

Simply type the IP address

preceded by ftp:// in the web
browser window.

ni.com

If you are unfamiliar with FTP commands and feel more comfortable using
a web browser, simply launch Internet Explorer and type the IP address to the
RT-PXI Controller.

LabVIEW Real-Time 6-10 © National Instruments Corporation

 Exercise 6.2

Students transfer files between host PC and

RT Series PXI Controller using FTP.

ni.com

© National Instruments Corporation 6-11 LabVIEW Real-Time

 Web Server (PXI and FP-2000 only)
The PXI and FP-2000 Controllers
also run a Web server, which can
be accessed with a web browser
like Internet Explorer.

While targeted to a PXI

or FP-2000 Controller, you
can enable the web server
by going to Tools »
Network Options.

ni.com

The RT-PXI also runs a Web server, which must be enabled through LabVIEW RT
while targeted to the RT-PXI Controller. Go to Tools»Network Options,
and check Enable Web Server. The default web server path is
c:\NI-RT\system\WWW. You can place image files, HTML files,
etc. in this directory on the RT-PXI and then browse to them remotely from
a web browser.

LabVIEW Real-Time 6-12 © National Instruments Corporation

 Web Server II
Once the PXI or FP-2000
Web server has been enabled,
you can browse .html files or
images across a network using
a web browser.

By saving data to an .html file

within LabVIEW Real-Times’
web server, you can publish
your data across the network
formatted in HTML.

ni.com

The web server is a powerful tool on the RT-PXI because you can incorporate
real-time control and file I/O to publish your results over the Web.

© National Instruments Corporation 6-13 LabVIEW Real-Time

 Exercise 6.3

Students browse RT Series Networked

Device’s Web server

ni.com

LabVIEW Real-Time 6-14 © National Instruments Corporation

 Watchdogs
!The RT Series PXI Controllers and FP-2000
Controllers have built-in watchdogs
• Programmable through watchdog VIs
• Can reset controller or set occurrence when
watchdog expires
• Can activate trigger on PXI back plane
!The FP-2000 is also equipped with a network
watchdog to respond to a connection failure
• Disabled by default
• Enable through FieldPoint Explorer

ni.com

The point to having a software watchdog is to disable a critical system in the event
of a software crash. The software watchdog on the PXI Controller is actually
a counter, which can trigger several different expiration actions if it expires,
including resetting the PXI Controller, setting an occurrence in software, and
activating a PXI back plane trigger.

© National Instruments Corporation 6-15 LabVIEW Real-Time

 Watchdog Timer

Expires when not

whacked in time.

The RT PXI and FP-2000

Controllers can respond to
an expired watchdog by:
• Resetting
• Setting a software occurrence
• Setting an external trigger line

ni.com

The watchdog API can be found under the RT functions palette. The top-level
watchdog API consists of three main VIs—Configure Watchdog, Whack Watchdog,
and Clear Watchdog.

LabVIEW Real-Time 6-16 © National Instruments Corporation

 Exercise 6.4

Students implement a watchdog timer on the

RT Series PXI Controller.

ni.com

© National Instruments Corporation 6-17 LabVIEW Real-Time

 Software Watchdogs II
First enable the network
watchdog on the network
module

ni.com

LabVIEW Real-Time 6-18 © National Instruments Corporation

 Software Watchdogs II
Next enable network watchdog
for each output module

ni.com

© National Instruments Corporation 6-19 LabVIEW Real-Time

 Software Watchdogs II
Finally, specify channels and
output response to the network
watchdog

ni.com

LabVIEW Real-Time 6-20 © National Instruments Corporation

 Other Types of Watchdogs
!Other software watchdogs
• Increment a byte of shared memory and monitor
“heartbeat” on host or RT PXI
• Use TCP/IP or VI Server to increment a variable
and monitor “heartbeat”
!Hardware watchdogs
• Toggle digital line on a DAQ device and monitor
line state on RT PXI or external device
• Perform single pulse generation with a counter
and reprogram counter within a loop
• Perform pattern generation on 653x device and
reset the buffer before it finishes
ni.com

In addition to the built-in watchdog available on the FP-2000 and RT PXI

Controllers, programmers can implement their own software watchdog using shared
memory (7030 only), TCP/IP, or VI Server. The RT examples have an example that
implements a software watchdog through shared memory on the 7030 boards.
Besides software watchdogs, users can also implement hardware watchdogs to alert
other hardware if your real-time application crashes or fails. Counters are especially
useful for watchdog applications because they behave very similar to PXI watchdog
and can be programmed similarly too. A less obvious watchdog solution involves
the 653x devices. By configuring a 653x device for pattern output, you can output
a buffer of known length at some user-defined frequency, and reset the buffer in
a software loop before the device reaches the end. If the device reaches the last
pattern, for example, switches from 0 to 255, we can deduce from the digital lines
on the 653x device that the software has failed to reset the buffer on time.

© National Instruments Corporation 6-21 LabVIEW Real-Time

 Exercise 6.5 (Optional)

Students implement a hardware watchdog

on an E-Series device and on a 653x device.

ni.com

LabVIEW Real-Time 6-22 © National Instruments Corporation

 Chapter 6 - Summary

!Applications can be embedded

or downloaded
!FTP Server and Web Server available
!Hardware and Software Watchdogs for safety

ni.com

© National Instruments Corporation 6-23 LabVIEW Real-Time

 Exercise 6-1
Objective: To build an RT executable that automatically starts itself after
power failure.

1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. Select File»Open and open MyApp VI from Exercises\Chapter 6.
3. Examine the diagram and run the VI on the target, watching the signal accessory.
Be sure to select the appropriate DAQ device number and make it the default
value. (Note: If you fail to make the DAQ device number the default value, later
steps will result in errors.)
4. Save the VI.

PART 1. (PXI only)

1. Close the VI from memory and open a new VI while still targeted to your
PXI Controller.
2. Select Tools»Network Options…
3. Select RT Target Miscellaneous and take note of Application Path, which
defaults to c:\ni-rt\startup\startup.exe.

4. Check the box labeled Launch Application at Boot-Up, and then click OK.
5. Next select Tools»Build Application and click on the Source Files tab.
6. Add MyApp VI as the top-level VI.

LabVIEW Real-Time 6-24  National Instruments Corporation

 7. Click on the Application Settings tab, and uncheck the option to Show
LabVIEW Real-Time target selection dialog box when launched.
8. Click Build, then click Done.
9. When prompted to create script file, choose No.

We have just compiled MyApp VI into an executable. Because we built

the executable while targeted to our PXI, the executable was downloaded
to c:\ni-rt\startup\startup.exe on the PXI’s hard disk. Next time
the PXI reboots, the executable will automatically begin running.

10. Simulate a power failure by recycling power on the PXI controller then monitor
the signal accessory as the PXI boots up.

 National Instruments Corporation 6-25 LabVIEW Real-Time

 PART 2. (For RT Series DAQ devices, although the exercise can be completed with
an RT Series PXI controller.)
1. Launch LabVIEW Real-Time targeted to the host computer.
2. Open a new VI and select Tools»Build Application.
3. On the Target tab, enter the following information in each field:

4. Then follow the remaining steps as above, adding MyApp VI as the top-level VI
and building the application.

This time we compiled the VI while targeted to the host computer. (This is needed
for RT Series DAQ Devices because the processor board has no hard drive on
which to store the executable.) We will need to use command-line arguments
to configure the executable to target itself to the RT device. We will also need
to put the executable in our Windows startup directory.

5. Once you have built the VI into an executable, it should be located on your
desktop, labeled MyApp.exe.
6. Right-click on it and choose Create Shortcut.
7. Right-click on the shortcut, and choose Properties.
8. In the field labeled Target, add a command-line argument to target the executable
to the 7030 device (i.e., -target DAQ::1, or -target [IP Address]).

LabVIEW Real-Time 6-26  National Instruments Corporation

 9. Next place the shortcut in the Window startup folder, which may be located
at C:\WINDOWS\All Users\Start Menu\Programs\StartUp.
10. Power off the desktop PC to simulate power loss and watch the signal accessory
(which should now be attached to your DAQ device) as the PC boots up again.

Note: For users of Windows NT, you can download a utility called TweakUI.exe from
Microsoft, which allows you to configure Windows NT to automatically log in upon
bootup. Also, the paths may be different for Windows NT.

 National Instruments Corporation 6-27 LabVIEW Real-Time

 Exercise 6-2
Objective: To become familiar with the LabVIEW Real-Time FTP server
on PXI controllers.

PART 1.
1. On your Windows PC, click Start»Run, then type ftp 130.164.48.22,
except use the IP address of your controller instead of 130.164.48.22.
2. When prompted for a username and password, just hit Enter.

3. Once you have reached the ftp command prompt, type “?” and then hit Enter.
4. After studying the list of ftp commands, try using the put and get commands
to transfer files back and forth. You may need to change local directory with lcd,
or change target directory with cd. You can list the contents of the remote
working directory with ls.

LabVIEW Real-Time 6-28  National Instruments Corporation

 PART 2.
1. Launch a Web browser, such as Internet Explorer.
2. Enter ftp://130.164.48.22 (use your own IP address here) in the address
field, and then hit Enter.

3. You should have immediate access to the remote hard drive via the Web browser,
so try dragging and dropping contents to transfer files.

Exercise 6-3
Objective: To become familiar with the LabVIEW Real-Time Web server
on PXI controllers.

1. Launch LabVIEW Real-Time and select New VI.

2. Select Tools»Network Options…
3. Select Web Server: Configuration and check the box labeled Enable Web
Server. Take note of the web server’s root directory, which can be modified.
4. Close LabVIEW Real-Time.
5. Locate image.gif under \Exercises\Chapter 6 on the host PC running
Windows.
6. Open an FTP connection to your PXI and transfer image.gif to
ftp://130.164.48.22/NI-RT/SYSATEM/WWW/

 National Instruments Corporation 6-29 LabVIEW Real-Time

 7. Open a web browser and navigate to
http://130.164.48.22/image.gif.

Exercise 6-4
Objective: To implement a watchdog timer, which will reboot the PXI
controller when the watchdog expires.

1. Complete Part 1 of Exercise 6-1 above.

2. Select File»Open and open Watchdog Reset – RT Engine VI from
c:\Program Files\National Instruments\LabVIEW
6\Examples\RT\RT Watchdog(PXI-8156B).llb.
3. The block diagram comprises 3 subVI nodes: Watchdog Configure, Watchdog
Whack, and Watchdog Clear. Watchdog Configure configures and arms
a hardware counter/timer built into the RT Series PXI Controller to count
down for a period equal to the timeout input parameter. Once the internal
counter has expired (i.e., counted past the timeout period), the Controller
responds according to the expiration actions. In this example, we will choose
to reset the PXI Controller.

LabVIEW Real-Time 6-30  National Instruments Corporation

 Watchdog Whack simply resets the watchdog timer back to zero each time
it is called. We can conclude then that by calling Watchdog Whack more
frequently than the period set up in Watchdog Configure, the watchdog
counter will not expire. The point to having a software watchdog is to disable
a critical system in the event of a software crash.

4. Study the front panel of Watchdog Reset – RT Engine VI. Notice the
watchdog timeout period is 0.25 seconds and the loop period is set for only
100 milliseconds or 0.10 seconds, which ensures that Watchdog Whack will
have plenty of time to reset the watchdog counter before it resets.

 National Instruments Corporation 6-31 LabVIEW Real-Time

 5. Download the VI to the RT PXI target and run it.
6. Slowly increase the loop cycle time by adjusting the Wait control
on the front panel.
7. Once you reach 250+ millisecond loop periods, notice the PXI Controller
reboots itself.
8. Also monitor the signal accessory, which should be blinking as it did
in Exercise 6-1 once the controller boots up again.

Exercise 6-5 (Optional)

Objective: To implement a hardware watchdog using pulse train
generation (Part 1) and pattern output (Part 2).

Note: This section requires knowledge of advanced DAQ programming and timing
concepts and a 653x board for PART 2.

PART 1.
1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. Select File»Open and open Watchdog Example VI from
Exercises\Chapter 6\counterWatchdog.llb.

LabVIEW Real-Time 6-32  National Instruments Corporation

 3. Study the diagram, which is divided into three simple parts.

4. The first subVI sets up a DAQ counter on either a 660x device or an E-Series
device to perform pulse generation.
5. The second subVI resets the counter within the WHILE loop.
6. Finally, the last subVI clears and disarms the DAQ counter.
This example is almost identical to Watchdog Reset – RT Engine VI because
both examples configure a counter to perform a task after counting down some
time period. The WHILE loops in both examples reprogram the counter to start
counting down again, and if the WHILE loops are not executed on time,
the counter performs the watchdog expiration task. In this case, however,
the watchdog merely outputs a pulse rather than resetting the PXI controller
or setting an occurrence.
7. Download and run Watchdog Example VI using a Watchdog Delay of 5 seconds
and a Loop Delay of 3 seconds.
8. Slowly increase the loop delay to 5+ seconds while monitoring the output
of counter 0 on the DAQ device.

PART 2.
1. Launch LabVIEW Real-Time and select the appropriate RT Series device from
the Select Target Platform dialog box. Close any VIs already running on the
target platform.
2. Select File»Open and open digWatchdog VI from Exercises\Chapter 6\
digWatchdog.llb.
3. Study the diagram, which contains mostly advanced digital I/O functions.

 National Instruments Corporation 6-33 LabVIEW Real-Time

 The example initializes a buffer of several 0’s followed by a 1. The example
programs a 653x device to output this buffer on one digital line at a user-define
rate. The WHILE loop simply programs the device to restart the output at the
beginning of the buffer. If the WHILE loop executes more frequently than device
can output one entire buffer, the device will never reach the last pattern in the
buffer (i.e., the board will never change the output state on line 0).

On the other hand, if the loop dies or slows down significantly, the 653x device
will reach the end of the buffer and output a 1 or logic high. This voltage change
could be tied to a relay or some other external circuit to shut down a critical
system in the event of a software failure.

4. Download and run the VI with a clock frequency of 10 Hz and a sleep time
of 9 seconds or 9000 ms.
5. While the VI is running, monitor the state of digital line 0 on the 653x device.
It should be 0 volts.
6. Now increase the sleep time to 10+ seconds. What happens to digital line 0?

LabVIEW Real-Time 6-34  National Instruments Corporation