When the Routines category is selected in the Explorer panel, the Report panel shows profiling results for individual routines. The panel contains several columns that do not depend on the active counter and there is also a set of columns that depend on the counter that was used during profiling.

Below is a description of the available columns:

Columns that do not depend on the active counter

Columns (in alphabetical order)	Description
Address	Routine’s address in memory. This column is used for unmanaged (native-code) routines only.
Analysis Result	Specifies if the routine was instrumented or not. If the routine was instrumented, this column is empty. Otherwise, the column displays a short description why the routine was not instrumented: Less than 5 bytes - The routine occupies less than 5 bytes in memory (for 32-bit applications). See Profiling Small Functions. Less than 6 bytes - The routine occupies less than 6 bytes in memory (for 64-bit applications). See Profiling Small Functions. No line info - The routine was added to a line-level area, but the debug information contains no info about routine lines. These routines can be profiled at routine level only. Unsafe code - AQTime was not able to instrument the routine safely. This typically occurs when the routine’s binary code is intermixed with data areas. See Profiling Routines That Have Unsafe Code. No ret instruction - The routine’s binary code does not contain the `ret` instruction (this may happen if the routine finishes with the `jmp` instruction). See Profiling Routines That Do Not Have the ret Instruction. Duplicated code - The routine whose code coincides with code of another routine. To learn more about this, see Profiling Duplicated Code.
Class Name	Name of the class where the method is defined.
Code Type	Specifies the type of the routine’s code. The following values are possible: MSIL - Managed-code routine with MSIL (Microsoft Intermediate Language)code. x64 - 64-bit code routine. x86 - Native-code (unmanaged) routine. Pseudo - Pseudo routine that was created by the context. For example, <JIT Compiler>, <Garbage Collector>, <Unknown PInvoke> or <Root>. PInvoke - Native-code routine for which there is a declaration in one of managed modules and that is called from within the unmanaged code. NGen - Managed routine that was compiled by the ngen utility (CLR Native Image Generator) with the /prof argument in its command line. The ngen compilation means the routine was compiled before the application starts. Script - The routine belongs to a script that was profiled along with the host application. See Profiling Scripts - Overview for details.
Exceptions (#)	Number of times the method was entered but not successfully exited. This is usually a count of exception exits.
Hit Count	The number of routine calls that were profiled. See also Skip Count. The total number of times the routine was executed is determined as Hit Count + Skip Count.
Max Recursion Depth	The maximum number of nested (recursive) calls to the function reached during the run.
Module Name	The name of the module which contains the profiled routine.
Namespace	Namespace of the method’s class (this column is used for managed routines only).
Routine Name	Method name.
Skip Count	Number of times the routine was excluded from profiling, because the profiling status was off (this can be, for example, the number of times the routine was affected by an off-trigger or the number of times the routine was executed when the Enable/Disable Profiling button was not pressed). See also Hit Count. The total number of times the routine was executed is determined as Hit Count + Skip Count.
Source File	Name of the source file for the method. The values for this column are read from the application’s debug info. If debug info does not contain information on the file name, the column is empty.
Source Line	Source file’s line number where the method’s implementation begins. The values for this column are read from the application’s debug info.
Token	The routine’s token.
Unit Name	The name of the linkage unit containing the routine. This column is used for unmanaged (native-code) routines only.

Columns specific to the Elapsed Time, User Time and User+Kernel Time counters

You can specify the measurement unit for the following columns (seconds, milliseconds, microseconds or machine cycles) using the Counter unit box on the Profiler toolbar. See also Profiler Toolbar - Performance Profiler Items.

Columns (in alphabetical order)

Description

Average Time

Average time spent executing the routine’s code on one call. This is simply Time / (Hit Count + Skip Count).

Average Time with Children

Average time spent on each call to the routine, child calls included. This is simply Time with Children / (Hit Count + Skip Count).

First Time

Time spent executing the routine’s code on the first call (child calls are excluded).

First Time With Children

Time spent executing the routine’s code on the first call (including time of child calls that were made during the first routine’s call).

Note:

The First Time and First Time With Children columns help you determine why a routine took too much time to execute: a routine itself can work quickly, but it can perform initialization actions on the first call. These actions may perform slowly and make the routine one of the slowest routines in your application. By comparing results in the First Time (First Time With Children) and Time (Time With Children) columns, you can determine where the bottleneck occurred.

Max Time and
Min Time

Maximum and minimum time spent executing the routine’s code on a call. Exceptional values point out perhaps unexpected special conditions.

Max Time with Children and
Min Time with Children

Maximum and minimum time spent executing the routine’s code on a call (including child function calls). Exceptional values point out perhaps unexpected special conditions.

Shared Time

Total time spent executing the routine’s code, as a percentage of the total time spent on calls to the routine including calls to child routines.This is simply (Time / Time with Children)*100.

Time

Total time spent executing the routine’s code excluding child calls. The sum of all profiled methods appears in the footer of this column.

Time with Children

Total time spent on calls to the routine including calls to child routines. The sum for all profiled routines is displayed in the footer of this column. It will normally be an important multiple of actual profile-run duration, as child calls are counted several times.

% Time

Total time spent executing the routine’s code, as a percentage of the time spent executing all profiled routines.

% with Children

Time with Children value as a percentage of the sum of Time with Children for all profiled routines.

Columns specific to the CPU Cache Misses counter

Columns (in alphabetical order)	Description
Average Misses	Average number of cache misses that occurred during execution of the method’s code on one call. This is simply Misses / (Hit Count + Skip Count).
Average Misses with Children	Average number of cache misses that occurred during the routine execution (including cache misses that occurred in child calls). This is simply Misses with Children / (Hit Count + Skip Count).
First Misses	The number of cache misses that occurred during execution of the function’s code on the first call (child calls are excluded).
First Misses With Children	The number of cache misses that occurred during execution of the function’s code on the first call (including child calls).
Max Misses and Min Misses	Maximum and minimum number of cache misses that occurred during execution of the method’s code on a call. Exceptional values point out perhaps unexpected special conditions.
Max Misses with Children and Min Misses with Children	Maximum and minimum number of cache misses that occurred during the routine execution (including child function calls). Exceptional values point out perhaps unexpected special conditions.
Misses	Total number of cache misses that occurred during execution of the routine’s code excluding child calls. The sum for all profiled routines appears in the footer of this column.
Misses with Children	Total number of cache misses that occurred during execution of the routine (including its calls to child methods). The sum for all profiled routines is displayed in the footer of this column.
Shared Misses	Total number of cache misses that occurred during the routine execution (excluding child calls), as a percentage of the total number of cache misses that occurred during the routine execution (including child calls). That is, this is simply (Misses / Misses with Children)*100.
% Misses	Total number of cache misses that occurred during the routine execution (excluding child calls), as a percentage of the sum of cache misses that occurred during execution of all profiled routines.
% with Children	Total number of cache misses as a percentage of the sum of Misses with Children for all profiled routines.

Columns specific to the CPU Mispredicted Branches counter

Columns (in alphabetical order)	Description
Average Branches	Average number of branches that mispredicted during execution of the method’s code on one call. This is simply Branches / (Hit Count + Skip Count).
Average Branches with Children	Average number of branches that were mispredicted during the routine execution (including branches mispredicted in child calls). This is simply Branches with Children / (Hit Count + Skip Count).
Branches	Total number of branches that were mispredicted during execution of the routine’s code excluding child calls. The sum for all profiled routines appears in the footer of this column.
Branches with Children	Total number of branches that were mispredicted during execution of the routine (including mispredictions in child methods). The sum for all profiled routines is displayed in the footer of this column.
First Branches	The number of mispredictions that occurred during execution of the function’s code on the first call (child calls are excluded).
First Branches With Children	The number of mispredictions that occurred during execution of the function’s code on the first call (including child calls).
Max Branches and Min Branches	Maximum and minimum number of mispredictions that occurred during execution of the method’s code on a call. Exceptional values point out perhaps unexpected special conditions.
Max Branches with Children and Min Branches with Children	Maximum and minimum number of mispredictions that occurred during the routine execution (including mispredictions in child functions). Exceptional values point out perhaps unexpected special conditions.
Shared Branches	Total number of mispredicted branches that occurred during the routine execution (excluding child calls), as a percentage of the total number of mispredictions that occurred during the routine execution (including child calls). That is, this is simply (Branches / Branches with Children)*100.
% Branches	Total number of branches that were mispredicted during the routine execution (excluding child calls), as a percentage of the sum of mispredictions occurred during execution of all profiled routines.
% with Children	Total number of mispredicted branches as a percentage of the sum of Branches with Children for all profiled routines.

Columns specific to the Hard Memory Page Faults, Soft Memory Page Faults and All Memory Page Faults counters

Columns (in alphabetical order)	Description
Average Faults	Average number of page faults that occurred during execution of the method’s code on one call. This is simply Faults / (Hit Count + Skip Count).
Average Faults with Children	Average number of page faults that occurred during the routine execution (including page faults that occurred in child calls). This is simply Faults with Children / (Hit Count + Skip Count).
Faults	Total number of page faults that occurred during execution of the routine’s code (child calls are excluded). The sum for all profiled routines appears in the footer of this column.
Faults with Children	Total number of page faults that occurred during execution of the routine (including page faults that occurred in child methods). The sum for all profiled routines is displayed in the footer of this column.
First Faults	The number of page faults that occurred during execution of the function’s code on the first call (child calls are excluded).
First Faults With Children	The number of page faults that occurred during execution of the function’s code on the first call (child calls are included).
Max Faults and Min Faults	Maximum and minimum number of page faults that occurred during execution of the method’s code on a call. Exceptional values point out perhaps unexpected special conditions.
Max Faults with Children and Min Faults with Children	Maximum and minimum number of page faults that occurred during the routine execution (including page faults that occurred in child functions). Exceptional values point out perhaps unexpected special conditions.
Shared Faults	Total number of page faults that occurred during the routine execution (excluding child calls), as a percentage of the total number of page faults that occurred during the routine execution (including child calls). That is, this is simply (Faults / Faults with Children)*100.
% Faults	Total number of page faults that occurred during the routine execution (excluding child calls), as a percentage of the sum of page faults that occurred during execution of all profiled routines.
% with Children	Total number of page faults as a percentage of the sum of Faults with Children for all profiled routines.

Columns specific to the Split Load Replays, Split Store Replays and Blocked Store Forwards Replays counters

Columns (in alphabetical order)	Description
Average Replays	Average number of replays that occurred during execution of the method’s code on one call. This is simply Replays / (Hit Count + Skip Count).
Average Replays with Children	Average number of replays that occurred during the routine execution (including replays that occurred in child calls). This is simply Replays with Children / (Hit Count + Skip Count).
First Replays	The number of replays that occurred during execution of the function’s code on the first call (child calls are excluded).
First Replays With Children	The number of replays that occurred during execution of the function’s code on the first call (child calls are included).
Max Replays and Min Replays	Maximum and minimum number of replays that occurred during execution of the method’s code on a call. Exceptional values point out perhaps unexpected special conditions.
Max Replays with Children and Min Replays with Children	Maximum and minimum number of replays that occurred during the routine execution (including replays that occurred in child functions). Exceptional values point out perhaps unexpected special conditions.
Replays	Total number of replays that occurred during execution of the routine’s code (child calls are excluded). The sum for all profiled routines appears in the footer of this column.
Replays with Children	Total number of replays that occurred during execution of the routine (including replays that occurred in child methods). The sum for all profiled routines is displayed in the footer of this column.
Shared Replays	Total number of replays that occurred during the routine execution (excluding child calls), as a percentage of the total number of replays that occurred during the routine execution (including child calls). That is, this is simply (Replays / Replays with Children)*100.
% Replays	Total number of replays that occurred during the routine execution (excluding child calls), as a percentage of the sum of replays that occurred during execution of all profiled routines.
% with Children	Total number of replays as a percentage of the sum of Replays with Children for all profiled routines.

Columns specific to the 64K Aliasing Conflicts counter

Columns (in alphabetical order)	Description
Average Conflicts	Average number of aliasing conflicts that occurred during execution of the method’s code on one call. This is simply Conflicts / (Hit Count + Skip Count) .
Average Conflicts with Children	Average number of aliasing conflicts that occurred during the routine execution (including conflicts that occurred in child calls). This is simply Conflicts with Children / (Hit Count + Skip Count).
Conflicts	Total number of aliasing conflicts that occurred during execution of the routine’s code (child calls are excluded). The sum for all profiled routines appears in the footer of this column.
Conflicts with Children	Total number of aliasing conflicts that occurred during execution of the routine (including conflicts that occurred in child methods). The sum for all profiled routines is displayed in the footer of this column.
First Conflicts	The number of conflicts that occurred during execution of the function’s code on the first call (child calls are excluded).
First Conflicts With Children	The number of aliasing conflicts that occurred during execution of the function’s code on the first call (child calls are included).
Max Conflicts and Min Conflicts	Maximum and minimum number of aliasing conflicts that occurred during execution of the method’s code on a call. Exceptional values point out perhaps unexpected special conditions.
Max Conflicts with Children and Min Conflicts with Children	Maximum and minimum number of aliasing conflicts that occurred during the routine execution (including conflicts that occurred in child functions). Exceptional values point out perhaps unexpected special conditions.
Shared Conflicts	Total number of aliasing conflicts that occurred during the routine execution (excluding child calls), as a percentage of the total number of aliasing conflicts that occurred during the routine execution (including child calls). That is, this is simply (Conflicts / Conflicts with Children)*100.
% Conflicts	Total number of aliasing conflicts that occurred during the routine execution (excluding child calls), as a percentage of the sum of conflicts that occurred during execution of all profiled routines.
% with Children	Total number of aliasing conflicts as a percentage of the sum of Conflicts with Children for all profiled routines.

Columns specific to the Context Switches counter

Columns (in alphabetical order)	Description
Average Switches	Average number of context switches that occurred during execution of the method’s code on one call. This is simply Switches / (Hit Count + Skip Count).
Average Switches with Children	Average number of context switches that occurred during the routine execution (including switches that occurred in child calls). This is simply Switches with Children / (Hit Count + Skip Count).
First Switches	The number of context switches that occurred during execution of the function’s code on the first call (child calls are excluded).
First Switches With Children	The number of context switches that occurred during execution of the function’s code on the first call (child calls are included).
Max Switches and Min Switches	Maximum and minimum number of context switches that occurred during execution of the method’s code on a call. Exceptional values point out perhaps unexpected special conditions.
Max Switches with Children and Min Switches with Children	Maximum and minimum number of context switches that occurred during the routine execution (including switches that occurred in child functions). Exceptional values point out perhaps unexpected special conditions.
Shared Switches	Total number of context switches that occurred during the routine execution (excluding child calls), as a percentage of the total number of context switches that occurred during the routine execution (including child calls). That is, this is simply (Switches / Switches with Children)*100.
Switches	Total number of context switches that occurred during execution of the routine’s code (child calls are excluded). The sum for all profiled routines appears in the footer of this column.
Switches with Children	Total number of context switches that occurred during execution of the routine (including switches that occurred in child methods). The sum for all profiled routines is displayed in the footer of this column.
% Switches	Total number of context switches that occurred during the routine execution (excluding child calls), as a percentage of the sum of switches that occurred during execution of all profiled routines.
% with Children	Total number of context switches as a percentage of the sum of Switches with Children for all profiled routines.

A replay is an attempt of executing a micro-operation
when conditions for the correct execution of this operation are not satisfied.
Replays may be caused by cache misses, store forwarding issues, etc.
Normally, certain number of replays always occur during the application
execution. However, a superfluous number of replays designates a
performance problem.

Routines Category

See Also