Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Totalview Examining Arrays

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

TotalView

Examining Arrays

version 8.6

Copyright 20072008 by TotalView Technologies. All rights reserved


Copyright 19982007 by Etnus LLC. All rights reserved.
Copyright 19961998 by Dolphin Interconnect Solutions, Inc.
Copyright 19931996 by BBN Systems and Technologies, a division of BBN Corporation.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without the prior written permission of TotalView Technologies.
Use, duplication, or disclosure by the Government is subject to restrictions as set forth in subparagraph (c)(1)(ii) of the Rights
in Technical Data and Computer Software clause at DFARS 252.227-7013.
TotalView Technologies has prepared this manual for the exclusive use of its customers, personnel, and licensees. The information in this manual is subject to change without notice, and should not be construed as a commitment by TotalView Technologies. TotalView Technologies assumes no responsibility for any errors that appear in this document.
TotalView and TotalView Technologies are registered trademarks of TotalView Technologies.
TotalView uses a modified version of the Microline widget library. Under the terms of its license, you are entitled to use these
modifications. The source code is available at http://www.totalviewtech.com/Products/TotalView/developers.
All other brand names are the trademarks of their respective holders.

Contents

Examining Arrays
Examining and Analyzing Arrays ......................................................................................... 1
Displaying Array Slices ................................................................................................. 2
Using Slices and Strides ............................................................................................ 2
Using Slices in the Lookup Variable Command ........................................................ 4
Array Slices and Array Sections .................................................................................... 4
Filtering Array Data Overview ....................................................................................... 5
Filtering by Comparison ............................................................................................ 6
Filtering for IEEE Values ............................................................................................ 7
Filtering a Range of Values ........................................................................................ 9
Creating Array Filter Expressions .............................................................................. 9
Using Filter Comparisons ........................................................................................ 10
Sorting Array Data ...................................................................................................... 10
Obtaining Array Statistics ........................................................................................... 11
Displaying a Variable in all Processes or Threads ............................................................. 13
Diving on a Show Across Pointer ............................................................................ 14
Editing a Show Across Variable ............................................................................... 14
Visualizing Array Data ....................................................................................................... 15
Visualizing a Show Across Variable Window ............................................................ 15

INDEX ........................................................................................................................................ 17

Examining Arrays: version 8.6

iii

iv

Contents

Examining Arrays

This chapter explains how to examine and change array data as you
debug your program. Since arrays also appear in the Variable Window, you need to be familiar with the information in Chapter 2,
Examining and Changing Data, on page v.
The topics in this chapter are:
Q
Q
Q

Examining and Analyzing Arrays on page 1


Displaying a Variable in all Processes or Threads on page 13
Visualizing Array Data on page 15

Examining and Analyzing Arrays


TotalView can quickly display very large arrays in Variable Windows. An array
can be the elements that you define in your program, or it can be an area
of memory that you cast into an array.
If an array extends beyond the memory section in which it resides, the initial portion of the array is formatted correctly. If memory isnt allocated for
an array element, TotalView displays Bad Address in the elements subscript.
Topics in this section are:
Q
Q
Q
Q

Displaying Array Slices on page 2


Filtering Array Data Overview on page 5
Sorting Array Data on page 10
Obtaining Array Statistics on page 11

Examining Arrays: version 8.6

Examining and Analyzing Arrays

Displaying Array Slices


TotalView lets you display array subsections by editing the Slice field in an
arrays Variable Window. (An array subsection is called a slice.) The Slice field
contains placeholders for all array dimensions. For example, the following
is a C declaration for a three-dimensional array:
integer an_array[10][20][5]

Because this is a three-dimensional array, the initial slice definition is


[:][:][:]. This lets you know that the array has three dimensions and that
TotalView is displaying all array elements.
The following is a deferred shape array definition for a two-dimensional
array variable:
integer, dimension (:,:) :: another_array

The TotalView slice definition is (:,:).


TotalView displays as many colons (:) as there are array dimensions. For
example, the slice definition for a one-dimensional array (a vector) is [:] for
C arrays and (:) for Fortran arrays.
CLI:

dprint -slice \[n:m\] an_array


dprint -slice (n:m,p:q) an_array

Using Slices and Strides


A slice has the following form:
lower_bound:upper_bound[:stride]
The stride, which is optional, tells TotalView to skip over elements and not
display them. Adding a stride to a slice tells the debugger to display every
stride element of the array, starting at the lower_bound and continuing through
the upper_bound, inclusive.
For example, a slice of [0:9:9] used on a ten-element C array tells TotalView
to display the first element and last element, which is the ninth element
beyond the lower bound.
If the stride is negative and the lower bound is greater than the upper
bound, TotalView displays a dimension with its indices reversed. That is,
TotalView treats the slice as if it was defined as follows:
[upperbound : lowerbound : stride]
CLI:

dprint an_array(n:m:p,q:r:s)

For example, the following definition tells TotalView to display an array


beginning at its last value and moving to its first:
[::-1]

Examining Arrays: version 8.6

Examining and Analyzing Arrays

This syntax differs from Fortran 90 syntax in that Fortran 90 requires that
you explicitly enter the upper and lower bounds when youre reversing the
order for displaying array elements.
Because the default value for the stride is 1, you can omit the stride (and
the colon that precedes it) from your definition. For example, the following
two definitions display array elements 0 through 9:
[0:9:1]
[0:9]

If the lower and upper bounds are the same, just use a single number. For
example, the following two definitions tell TotalView to display array element 9:
[9:9:1]
[9]

The lower_bound, upper_bound, and stride must be constants. They cannot be expressions.
Example 1

A slice declaration of [::2] for a C or C++ array (with a default lower bound
of 0) tells TotalView to display elements with even indices of the array; that
is, 0, 2, 4, and so on. However, if this were defined for a Fortran array
(where the default lower bound is 1), TotalView displays elements with odd
indices of the array; that is, 1, 3, 5, and so on.

Example 2

The following figure displays a stride of (::9,::9). This definition displays the
four corners of a ten-element by ten-element Fortran array.

Figure 1: Stride Displaying


the Four Corners of an
Array

Example 3

You can use a stride to invert the order and skip elements. For example, the
following slice begins with the upper bound of the array and displays every
other element until it reaches the lower bound of the array:
(::2)

Using (::2) with a Fortran integer(10) array tells TotalView to display the
elements 10, 8, 6, 4, and 2.
Example 4

You can simultaneously invert the arrays order and limit its extent to display a small section of a large array. The following figure shows how to

Examining Arrays: version 8.6

Examining and Analyzing Arrays

specify a (2:3,7::1) slice with an integer*4(1:5,2:10) Fortran array. (See


Figure 2 on page 4.)
Figure 2: Fortran Array with
Inverse Order and Limited
Extent

After you enter this slice value, TotalView only shows elements in rows 2
and 3 of the array, beginning with column 10 and ending with column 7.

Using Slices in the Lookup Variable Command


When you use the View > Lookup Variable command to display a Variable
Window, you can include a slice expression as part of the variable name.
Specifically, if you type an array name followed by a set of slice descriptions
in the View > Lookup Variable command dialog box, TotalView initializes
the Slice field in the Variable Window to this slice description.
If you add subscripts to an array name in the View > Lookup Variable dialog box, TotalView will look up just that array element.
CLI:

dprint small_array(5,5)

You can, of course, type an expression into the View > Lookup Variable
dialog box; for example, you could type small_array(i-1,j-1).

Array Slices and Array Sections


An array slice allows you to see a part of an array. The slice allows you to
remove parts of the array you do not want to see. For example, if you have
a 10,000 element array, you could tell TotalView that it should only display
100 of these elements. Fortran has introduced the concept of an array section. When you create an array section, you are creating a new array that is
a subset of the old array. Because it is a new array, its first array index is 1.
In Figure 3 on page 5, the top left Variable Window displays an eleven-element array slice. The bottom right Variable Window displays an eleven-element array.

Examining Arrays: version 8.6

Examining and Analyzing Arrays

Figure 3: An Array Slice and


an Array Section

While the data in both is identical, notice that the array numbering is different. In addition, the array slice shows an address for the array. The section,
however, only exists within TotalView. Consequently, there is no address
associated with it.

Filtering Array Data Overview


You can restrict what TotalView displays in a Variable Window by adding a
filter to the window. You can filter arrays of type character, integer, or floating point. Your filtering options are:
Q
Q
Q
Q

Arithmetic comparison to a constant value


Equal or not equal comparison to IEEE NaNs, Infs, and Denorms
Within a range of values, inclusive or exclusive
General expressions

When an element of an array matches the filter expression, TotalView


includes the element in the Variable Window display.
The following topics describe filtering options:
Q
Q
Q
Q
Q
Q

Filtering Array Data on page 6


Filtering by Comparison on page 6
Filtering for IEEE Values on page 7
Filtering a Range of Values on page 9
Creating Array Filter Expressions on page 9
Using Filter Comparisons on page 10

Examining Arrays: version 8.6

Examining and Analyzing Arrays

Filtering Array Data:

The procedure for filtering an array is simple:


select the Filter field, enter the array filter expression, and then press Enter.
TotalView updates the Variable Window to exclude elements that do not
match the filter expression. TotalView only displays an element if its value
matches the filter expression and the slice operation.
If necessary, TotalView converts the array element before evaluating the filter expression. The following conversion rules apply:

Q
Q

If the filter operand or array element type is floating point, TotalView converts the operand to a double-precision floating-point value. TotalView
truncates extended-precision values to double precision. Converting integer or unsigned integer values to double-precision values might result in a
loss of precision. TotalView converts unsigned integer values to nonnegative double-precision values.
If the filter operand or the array element is an unsigned integer, TotalView
converts the operand to an unsigned 64-bit integer.
If both the filter operand and array element are of type integer, TotalView
converts the values to type 64-bit integer.

TotalView conversion operations modify a copy of the arrays elements


conversions never alter the actual array elements.
To stop filtering an array, delete the contents of the Filter field in the Variable Window and press Enter. TotalView then updates the Variable Window
so that it includes all elements.

Filtering by Comparison
The simplest filters are ones whose formats are as follows:
operator value
where operator is either a C/C++ or Fortran-style comparison operator, and
value is a signed or unsigned integer constant or a floating-point number.
For example, the filter for displaying all values greater than 100 is:
> 100

The following table lists the comparison operators:


Comparison
Equal
Not equal
Less than
Less than or equal
Greater than
Greater than or equal

C/C++ Operator
==
!=
<
<=
>
>=

Fortran Operator
.eq.
.ne.
.lt.
.le.
.gt.
.ge.

The following figure shows an array whose filter is < 0. This tells TotalView
to only display array elements whose value is less than 0 (zero). See
Figure 4 on page 7.

Examining Arrays: version 8.6

Examining and Analyzing Arrays

Figure 4: Array Data Filtering


by Comparison

If the value youre using in the comparison is an integer constant, TotalView


performs a signed comparison. If you add the letter u or U to the constant,
TotalView performs an unsigned comparison.

Filtering for IEEE Values


You can filter IEEE NaN, Infinity, or denormalized floating-point values by
specifying a filter in the following form:
operator ieee-tag
The only comparison operators you can use are equal and not equal.
The ieee-tag represents an encoding of IEEE floating-point values, as the following table describes:
IEEE Tag Value
$nan
$nanq
$nans
$inf
$pinf
$ninf
$denorm
$pdenorm
$ndenorm

Meaning
NaN (Not a number), either quiet or signaling
Quiet NaN
Signaling NaN
Infinity, either positive or negative
Positive Infinity
Negative Infinity
Denormalized number, either positive or negative
Positive denormalized number
Negative denormalized number

Figure 5 on page 8 shows an example of filtering an array for IEEE values.


The bottom window in this figure shows how TotalView displays the unfiltered array. Notice the NaNQ, and NaNS, INF, and INF values. The other
two windows show filtered displays: the top window shows only infinite values; the remaining window only shows the values of denormalized numbers.

Examining Arrays: version 8.6

Examining and Analyzing Arrays

Figure 5: Array Data Filtering


for IEEE Values

If you are writing an expression, you can use the following Boolean functions to check for a particular type of value:
IEEE Intrinsic
$is_denorm(value)
$is_finite(value)
$is_inf(value)
$is_nan(value)
$is_ndenorm(value)
$is_ninf(value)
$is_nnorm(value)
$is_norm(value)
$is_nzero(value)
$is_pdenorm(value)
$is_pinf(value)
$is_pnorm(value)
$is_pzero(value)
$is_qnan(value)
$is_snan(value)
$is_zero(value)

Meaning
Is a denormalized number, either positive or negative
Is finite
Is Infinity, either positive or negative
Is a NaN (Not a number), either quiet or signaling
Is a negative denormalized number
Is negative Infinity
Is a negative normalized number
Is a normalized number, either positive or negative
Is negative zero
Is a positive denormalized number
Is positive Infinity
Is a positive normalized number
Is positive zero
Is a quiet NaN
Is a signaling NaN
Is zero, either positive or negative

Examining Arrays: version 8.6

Examining and Analyzing Arrays

Filtering a Range of Values


You can also filter array values by specifying a range, as follows:
[>] low-value : [<] high-value
where low-value specifies the lowest value to include, and high-value specifies
the highest value to include, separated by a colon. The high and low values
are inclusive unless you use less-than (<) and greater-than (>) symbols. If
you specify a > before low-value, the low value is exclusive. Similarly, a <
before high-value makes it exclusive.
The values of low-value and high-value must be constants of type integer,
unsigned integer, or floating point. The data type of low-value must be the
same as the type of high-value, and low-value must be less than high-value. If
low-value and high-value are integer constants, you can append the letter u
or U to the value to force an unsigned comparison. The following figure
shows a filter that tells TotalView that to only display values greater than
63, but less than 512. (See Figure 6.)
Figure 6: Array Data Filtering
by Range of Values

Creating Array Filter Expressions


The filtering capabilities described in the previous sections are those that
you use most often. In some circumstances, you may need to create a
more general expression. When you create a filter expression, youre creating a Fortran or C Boolean expression that TotalView evaluates for every
element in the array or the array slice. For example, the following expression displays all array elements whose contents are greater than 0 and less
than 50, or greater than 100 and less than 150:
($value > 0 && $value < 50) ||
($value > 100 && $value < 150)

Heres the Fortran equivalent:


($value .gt. 0 && $value .lt. 50) .or.
($value .gt. 100 .and. $value .lt.150)

The $value variable is a special TotalView variable that represents the current
array element. You can use this value when creating expressions.
Examining Arrays: version 8.6

Examining and Analyzing Arrays

Notice how the and and or operators are used in these expressions. The
way in which TotalView computes the results of an expression is identical
to the way it computes values at an eval point. For more information, see
Defining Eval Points and Conditional Breakpoints on page 18.

Using Filter Comparisons


TotalView provides several different ways to filter array information. For
example, the following two filters display the same array items:
> 100
$value > 100

The following filters display the same array items:


>0:<100
$value > 0 && $value < 100

The only difference is that the first method is easier to type than the second, so youre more likely to use the second method when youre creating
more complicated expressions.

Sorting Array Data


TotalView lets you sort the displayed array data into ascending or descending order. (It does not sort the actual data.) To sort (or remove the sort),
click the Value label.
Q
Q
Q

The first time you click, TotalView sorts the arrays values into ascending
order.
The next time you click on the header, TotalView reverses the order, sorting
the arrays values into descending order.
If you click again on the header, TotalView returns the array to its unsorted
order.

Here is an example that sorts an array into descending order:


Figure 7: Sorted Variable
Window

When you sort an arrays values, you are just rearranging the information
thats displayed in the Variable Window. Sorting does not change the order
10

Examining Arrays: version 8.6

Examining and Analyzing Arrays

in which values are stored in memory. If you alter what TotalView is displaying by using a filter or a slice, TotalView just sorts the values that could be
displayed; it doesnt sort all of the array.
If you are displaying the array created by a Show across commandsee
Displaying a Variable in all Processes or Threads on page 13 for more informationyou can sort your information by process or thread.

Obtaining Array Statistics


The Tools > Statistics command displays a window that contains information about your array. Figure 8 shows an example.
Figure 8: Array Statistics
Window

If you have added a filter or a slice, these statistics describe only the information currently being displayed; they do not describe the entire unfiltered
array. For example, if 90% of an arrays values are less than 0 and you filter
the array to show only values greater than 0, the median value is positive
even though the arrays real median value is less than 0.
TotalView displays the following statistics:
Q

Checksum
A checksum value for the array elements.
Count
The total number of displayed array values. If youre displaying a floatingpoint array, this number doesnt include NaN or Infinity values.
Denormalized Count
A count of the number of denormalized values found in a floating-point
array. This includes both negative and positive denormalized values as
defined in the IEEE floating-point standard. Unlike other floating-point
statistics, these elements participate in the statistical calculations.

Examining Arrays: version 8.6

11

Examining and Analyzing Arrays

12

Infinity Count
A count of the number of infinity values found in a floating-point array.
This includes both negative and positive infinity as defined in the IEEE
floating-point standard. These elements dont participate in statistical
calculations.
Lower Adjacent
This value provides an estimate of the lower limit of the distribution. Values below this limit are called outliers. The lower adjacent value is the first
quartile value minus the value of 1.5 times the difference between the first
and third quartiles.
Maximum
The largest array value.
Mean
The average value of array elements.
Median
The middle value. Half of the arrays values are less than the median, and
half are greater than the median.
Minimum
The smallest array value.
NaN Count
A count of the number of NaN (not a number) values found in a floatingpoint array. This includes both signaling and quiet NaNs as defined in the
IEEE floating-point standard. These elements dont participate in statistical calculations.
Quartiles, First and Third
Either the 25th or 75th percentile values. The first quartile value means
that 25% of the arrays values are less than this value and 75% are greater
than this value. In contrast, the third quartile value means that 75% of the
arrays values are less than this value and 25% are greater.
Standard Deviation
The standard deviation for the arrays values.
Sum
The sum of all the displayed arrays values.
Upper Adjacent
This value provides an estimate of the upper limit of the distribution. Values above this limit are called outliers. The upper adjacent value is the third
quartile value plus the value of 1.5 times the difference between the first
and third quartiles.
Zero Count
The number of elements whose value is 0.

Examining Arrays: version 8.6

Displaying a Variable in all Processes or Threads

Displaying a Variable in all Processes


or Threads
When youre debugging a parallel program that is running many instances
of the same executable, you usually need to view or update the value of a
variable in all of the processes or threads at once.
Before displaying a variables value in all threads or processes, you must
display an instance of the variable in a Variable Window. After TotalView displays this window, use one of the following commands:
Q

View > Show Across > Process, displays the value of the variable in all

processes.
Q
Q

View > Show Across > Thread, displays the value of a variable in all
threads within a single process.
View > Show Across > None, returns the window to what it was before
you used other Show Across commands.

You cannot simultaneously Show Across processes and threads in the same Variable
Window.
After using one of these commands, the Variable Window switches to an
array-like display of information, and displays the value of the variable in
each process or thread. Figure 9 shows a simple, scalar variable in each of
the processes in an OpenMP program.
Figure 9: Viewing Across
Threads

When looking for a matching stack frame, TotalView matches frames starting from the top frame, and considers calls from different memory or stack
locations to be different calls. For example, the following definition of

Examining Arrays: version 8.6

13

Displaying a Variable in all Processes or Threads


recurse() contains two additional calls to recurse(). Each of these calls gen-

erates a nonmatching call frame.


void recurse(int i) {
if (i <= 0)
return;
if (i & 1)
recurse(i 1);
else
recurse(i 1);
}

If the variables are at different addresses in the different processes or


threads, the field to the left of the Address field displays Multiple, and the
unique addresses appear with each data item.
TotalView also lets you Show Across arrays and structures. When you Show
Across an array, TotalView displays each element in the array across all processes. You can use a slice to select elements to be displayed in an across
display. The following figure shows the result of applying a Show Across >
Processes command to an array of structures. (See Figure 10.)
Figure 10: Viewing across an
Array of Structures

Diving on a Show Across Pointer


You can dive through pointers in a Show Across display. This dive applies to
the associated pointer in each process or thread.

Editing a Show Across Variable


If you edit a value in a Show Across display, TotalView asks if it should
apply this change to all processes or threads or only the one in which you
made a change. This is an easy way to update a variable in all processes.

14

Examining Arrays: version 8.6

Visualizing Array Data

Visualizing Array Data


The Visualizer lets you create graphical images of array data. This presentation lets you see your data in one glance and can help you quickly find
problems with your data while you are debugging your programs.
You can execute the Visualizer from within TotalView, or you can run it from
the command line to visualize data dumped to a file in a previous TotalView
session.
For information about running the Visualizer, see Chapter 2, Visualizing Programs and Data, on page 1.

Visualizing a Show Across Variable Window


You can export data created by using a Show Across command to the Visualizer by using the Tools > Visualize command. When visualizing this kind of
data, the process (or thread) index is the first axis of the visualization. This
means that you must use one less data dimension than you normally
would. If you do not want the process/thread axis to be significant, you can
use a normal Variable Window, since all of the data must be in one process.

Examining Arrays: version 8.6

15

Visualizing Array Data

16

Examining Arrays: version 8.6

Index

Symbols

array data filtering 5


checksum statistic 11
colon separators 2
count statistic 11
deferred shape 2
denormalized count statistic 11
displaying 1
displaying slices 2
filter conversion rules 6
filtering 5, 6
filtering expressions 9
filtering options 5
infinity count statistic 12
limiting display 3
lower adjacent statistic 12
lower bound of slices 2
maximum statistic 12
mean statistic 12
median statistic 12
minimum statistic 12
NaN statistic 12
overlapping nonexistent memory 1
quartiles statistic 12
skipping elements 3
slice example 2, 3
slices with the variable command
4
sorting 10
standard deviation statistic 12
statistics 11
stride 2
stride elements 2
subsections 2
sum statistic 12
upper adjacent statistic 12
upper bound of slices 2
viewing across elements 14
zero count statistic 12

$denorm filter 7
$inf filter 7
$is_denorm intrinsic function 8
$is_finite intrinsic function 8
$is_inf intrinsic function 8
$is_nan intrinsic function 8
$is_ndenorm intrinsic function 8
$is_ninf intrinsic function 8
$is_nnorm intrinsic function 8
$is_norm intrinsic function 8
$is_pdenorm intrinsic function 8
$is_pinf intrinsic function 8
$is_pnorm intrinsic function 8
$is_pzero intrinsic function 8
$is_qnan intrinsic function 8
$is_snan intrinsic function 8
$is_zero intrinsic function 8
$nan filter 7
$nanq filter 7
$nans filter 7
$ndenorm filter 7
$ninf filter 7
$pdenorm filter 7
$pinf filter 7
: as array separator 2

A
Array Data Filter by Range of Values
figure 9
array data filtering
by comparison 5
by range of values 9
for IEEE values 7
see also arrays, filtering
Array Data Filtering by Comparison figure 7
Array Data Filtering for IEEE Values figure 8
Array Statistics Window figure 11
arrays

Examining Arrays: version 8.6

C
C/C++

filter expression 9
checksum array statistic 11
CLI commands
dprint 2, 4
closed loop, see closed loop
colons as array separators 2
commands
Tools > Statistics 11
View > Lookup Variable 4
View > View Across > None 13
View > View Across > Process 13
View > View Across > Thread 13
comparisons in filters 10
conversion rules for filters 6
count array statistic 11

D
data filtering, see arrays, filtering
deferred shape array
definition 2
denorm filter 7
denormalized count array statistic 11
DENORMs 5
displaying
arrays 1, 2
diving
in a "view acrosss" pane 14
dprint command 2, 4

E
editing
view across data 14

F
figures
Array Data Filter by Range of Values 9
Array Data Filtering by Comparison 7
Array Data Filtering for IEEE Values
8

17

I
Array Statistics Window 11
Fortran Array with Inverse Order
and Limited Extent 4
Sorted Variable Window 10
Stride Displaying the Four Corners
of an Array 3
Viewing Across an Array of Structures 14
Viewing Across Threads 13
filter expression, matching 5
filtering
array data 5, 6
array expressions 9
by comparison 6
comparison operators 6
conversion rules 6
example 6
IEEE values 7
options 5
ranges of values 9
unsigned comparisons 7
filters 10
$denorm 7
$inf 7
$nan 7
$nanq 7
$nans 7
$ninf 7
$pdenorm 7
$pinf 7
comparisons 10
Fortran
filter expression 9
Fortran Array with Inverse Order and
Limited Extent figure 4
functions
IEEE 8

I
IEEE functions 8
inf filter 7
infinite loop, see loop, infinite
infinity count array statistic 12
INFs 5
intrinsic functions
$is_Inf 8
$is_inf 8
$is_nan 8
$is_ndenorm 8
$is_ninf 8
$is_nnorm 8
$is_norm 8
$is_pdenorm 8
$is_pinf 8
$is_pnom 8
$is_pzero 8
$is_qnan 8
$is_snan 8
$is_zero 8
inverting array order 3

18

L
limiting array display 3
Lookup Variable command
specifying slices 4
loop infinite, see infinite loop
lower adjacent array statistic 12
lower bounds
of array slices 2

M
matching stack frames 13
maximum array statistic 12
mean array statistic 12
median array statistic 12
minimum array statistic 12
Multiple indicator 14

N
nan filter 7
nanq filter 7
NaNs 5, 7
array statistic 12
nans filter 7
ndenorm filter 7
ninf filter 7
None (lView Across) command 13

O
omitting array stride 3
outliers 12

P
pdenorm filter 7
pinf filter 7
pthreads, see threads

Q
quartiles array statistic 12

S
skipping elements 3
slices
defining 2
descriptions 4
examples 2, 3
lower bound 2
of arrays 2
stride elements 2
upper bound 2
with the variable command 4
Sorted Variable Window figure 10
sorting
array data 10
stack frame
matching 13
standard deviation array statistic 12
statistics for arrays 11
stride 2
default value of 3
elements 2
in array slices 2

omitting 3
Stride Displaying the Four Corners of
an Array figure 3
structures
viewing across 14
sum array statistic 12

T
testing for IEEE values 8
Tools > Statistics command 11
Tools > Visualize command 15

U
upper adjacent array statistic 12
upper bounds
of array slices 2

V
Variable Window
lView Across display 13
view across 14
variables
at different addresses 14
View Across display 13
View > Lookup Variable command
specifying slices 4
View > View Across > None command 13
View Across
arrays and structures 14
view across
editing data 14
View Across None command 13
viewing across
variables 13
Viewing Across an Array of Structures
figure 14
Viewing Across Threads figure 13
Viewing Across Variable Window 14
viewing across variables and processes 13
viewing acrosscross
diving in pane 14
Visualize command 15
Visualizer 15

Z
zero count array statistic 12

Examining Arrays: version 8.6

You might also like