Leistung

Profilieren der Ausführungszeit, Profilieren der Speicherauslastung, optimierter Code

Wenn Sie bereit sind, Produktionscode zu generieren, können Sie Konfigurationsoptionen und erweiterte Optimierungen nutzen, um die Leistung zu verbessern. Sie können die in der Produktdokumentation von MATLAB^® Coder™ beschriebenen Optionen und Optimierungen verwenden. Mit Embedded Coder^® können Sie die Leistung zudem mithilfe von Ausführungszeit-Profilierung und Speicherauslastungsprofilierung analysieren.

Mit der Profilierung der Ausführungszeit können Sie:

Ermitteln, ob der generierte Code die Echtzeitanforderungen Ihrer Hardware erfüllt.
Herausfinden, bei welchen Codeabschnitten Leistungsverbesserungen erforderlich sind.

Die statischen Codemetrik-Berichte umfassen Metriken zu Dateien, globalen Variablen und Funktionen. Mit dem statischen Codemetrik-Bericht können Sie ermitteln, welche globalen Variablen und Funktions-Aufrufspfade sich auf die Leistung auswirken. Um die Größe des Stack-Speichers zu ermitteln, der zur Ausführung des generierten Codes erforderlich ist, können Sie eine Software-in-the-Loop- (SIL) und Processor-in-the-Loop-Ausführung (PIL) ausführen, die ein Stack-Auslastungsprofil ausgibt.

Apps

Code Profile Analyzer

Analyze execution-time and stack usage profiles for generated code (Seit R2023a)

Funktionen

alle erweitern

Hinzufügen von Code-Instrumentierung

instrumentCode Add instrumentation to generated code to perform execution time / memory usage profiling and analyze code coverage (Seit R2023a)

Ausführungszeit-Profilierung

`ExecutionTimeInSeconds`	Get execution time in seconds for profiled section of code (MATLAB code generation)
`ExecutionTimeInTicks`	Get execution times in timer ticks for profiled section of code (MATLAB code generation)
`getCoderExecutionProfile`	Extract execution-time profile for code generated from MATLAB function (MATLAB code generation)
`Name`	Get name of profiled code section (MATLAB code generation)
`Number`	Get number that uniquely identifies profiled code section (MATLAB code generation)
`NumCalls`	Total number of calls to profiled code section (MATLAB code generation)
`report`	Open code execution profiling report and specify display of time measurements (MATLAB code generation)
`schedule`	Visualize task scheduling (MATLAB code generation) (Seit R2021b)
`Sections`	Get array of `coder.profile.ExecutionTimeSection` objects for profiled code sections (MATLAB code generation)
`SelfTimeInTicks`	Get number of timer ticks recorded for profiled code section, excluding time spent in child functions (MATLAB code generation)
`Time`	Time over which code section execution time measurements are made (MATLAB code generation)
`TimerTicksPerSecond`	Get and set number of timer ticks per second (MATLAB code generation)
`TotalExecutionTimeInTicks`	Get total number of timer ticks recorded for profiled code section (MATLAB code generation)
`TotalSelfTimeInTicks`	Get total number of timer ticks recorded for profiled code section, excluding time spent in child functions (MATLAB code generation)
`TotalTurnaroundTimeInTicks`	Get total number of timer ticks between start and finish of the profiled code section over the entire execution. (MATLAB code generation)
`TurnaroundTimeInTicks`	Get number of timer ticks between start and finish of the profiled code section (MATLAB code generation)

Ausführungszeit-Profilierung – Höchstwerte

`MaximumExecutionTimeCallNum`	Get the call number at which maximum number of timer ticks occurred (MATLAB code generation)
`MaximumExecutionTimeInTicks`	Get maximum number of timer ticks for single invocation of profiled code section (MATLAB code generation)
`MaximumSelfTimeCallNum`	Get the call number at which the maximum number of timer ticks occurred, excluding time spent in child functions (MATLAB code generation)
`MaximumSelfTimeInTicks`	Get the maximum number of timer ticks recorded for profiled code section, excluding time spent in child functions (MATLAB code generation)
`MaximumTurnaroundTimeCallNum`	Get call number for the code section invocation with the maximum number of timer ticks between the start and the finish (MATLAB code generation)
`MaximumTurnaroundTimeInTicks`	Get maximum number of timer ticks between start and finish of a single invocation of profiled code section (MATLAB code generation)

Objekte

alle erweitern

Stackauslastungs-Profilierung

coder.profile.StackDriver Specify driver to obtain stack usage data from target hardware (Seit R2022a)

Themen

Ausführungszeit-Profilierung

Execution Time Profiling for SIL and PIL
Why measure execution times for code generated from entry-point functions.
Generate Execution Time Profile
Enable execution-time profiling for a software-in-the-loop (SIL) or processor-in-the-loop (PIL) execution.
View Execution Times
Open the code execution profiling report generated by a SIL or PIL execution.
Analyze Execution Time Data
Use line commands to analyze execution-time measurements from a SIL or PIL execution.
Investigate Execution-Time Issues Using PMU Metrics
Analyze code execution metrics from the Performance Monitoring Unit of the hardware.

Speicherauslastungsprofilierung

Generating a Static Code Metrics Report for Code Generated from MATLAB Code
Create and explore an example static code metrics report.
Static Code Metrics
The code generator performs static analysis of the generated C or C++ code and provides these metrics in the static code metrics report in the code generation report.
Stack Usage Profiling for Code Generated From MATLAB Code
Determine size of stack memory that is required to run generated code.
Preserve Unused Class Properties in Generated C/C++ Code
Preserve unused class properties or structure fields in the generated C/C++ code.

Ausführungsgeschwindigkeit

Simplify Multiply Operations for Array Indexing in Loops
Replace multiply operations with add operations in array indexing in loops in C/C++ code generated from MATLAB code.
Generate SIMD Code from MATLAB Functions for Intel Platforms
Improve the execution speed of the generated code using Intel^® SSE and Intel AVX technology.
Generate SIMD Code from MATLAB Functions for Apple silicon Platforms
Improve the execution speed of the generated code for Apple silicon platforms using ARM^® Neon technology.