QOpenGLTimeMonitor#
The QOpenGLTimeMonitor
class wraps a sequence of OpenGL timer query objects. More…
Synopsis#
Functions#
def
create
()def
destroy
()def
isCreated
()def
isResultAvailable
()def
objectIds
()def
recordSample
()def
reset
()def
sampleCount
()def
setSampleCount
(sampleCount)def
waitForIntervals
()def
waitForSamples
()
Note
This documentation may contain snippets that were automatically translated from C++ to Python. We always welcome contributions to the snippet translation. If you see an issue with the translation, you can also let us know by creating a ticket on https:/bugreports.qt.io/projects/PYSIDE
Detailed Description#
The QOpenGLTimeMonitor
class is a convenience wrapper around a collection of OpenGL timer query objects used to measure intervals of time on the GPU to the level of granularity required by your rendering application.
The OpenGL timer queries objects are queried in sequence to record the GPU timestamps at positions of interest in your rendering code. Once the results for all issues timer queries become available, the results can be fetched and QOpenGLTimerMonitor will calculate the recorded time intervals for you.
The typical use case of this class is to either profile your application’s rendering algorithms or to adjust those algorithms in real-time for dynamic performance/quality balancing.
Prior to using QOpenGLTimeMonitor
in your rendering function you should set the required number of sample points that you wish to record by calling setSamples(). Note that measuring N sample points will produce N-1 time intervals. Once you have set the number of sample points, call the create()
function with a valid current OpenGL context to create the necessary query timer objects. These steps are usually performed just once in an initialization function.
Use the recordSample()
function to delimit blocks of code containing OpenGL commands that you wish to time. You can check availability of the resulting time samples and time intervals with isResultAvailable()
. The calculated time intervals and the raw timestamp samples can be retrieved with the blocking waitForIntervals()
and waitForSamples()
functions respectively.
After retrieving the results and before starting a new round of taking samples (for example, in the next frame) be sure to call the reset()
function which will clear the cached results and reset the timer index back to the first timer object.
See also
- class PySide6.QtOpenGL.QOpenGLTimeMonitor([parent=None])#
- Parameters:
parent –
PySide6.QtCore.QObject
Creates a QOpenGLTimeMonitor
instance with the given parent
. You must call create()
with a valid OpenGL context before using.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.create()#
- Return type:
bool
Instantiate sampleCount()
OpenGL timer query objects that will be used to track the amount of time taken to execute OpenGL commands between successive calls to recordSample()
.
Returns true
if the OpenGL timer query objects could be created.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.destroy()#
Destroys any OpenGL timer query objects used within this instance.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.isCreated()#
- Return type:
bool
Returns true
if the underlying OpenGL query objects have been created. If this returns true
and the associated OpenGL context is current, then you are able to record time samples with this object.
- PySide6.QtOpenGL.QOpenGLTimeMonitor.isResultAvailable()#
- Return type:
bool
Returns true
if the OpenGL timer query results are available.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.objectIds()#
- Return type:
Returns a QList
containing the object Ids of the OpenGL timer query objects.
- PySide6.QtOpenGL.QOpenGLTimeMonitor.recordSample()#
- Return type:
int
Issues an OpenGL timer query at this point in the OpenGL command queue. Calling this function in a sequence in your application’s rendering function, will build up details of the GPU time taken to execute the OpenGL commands between successive calls to this function.
- PySide6.QtOpenGL.QOpenGLTimeMonitor.reset()#
Resets the time monitor ready for use in another frame of rendering. Call this once you have obtained the previous results and before calling recordSample()
for the first time on the next frame.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.sampleCount()#
- Return type:
int
Returns the number of sample points that have been requested with setSampleCount()
. If create was successfully called following setSampleCount()
, then the value returned will be the actual number of sample points that can be used.
The default value for sample count is 2, leading to the measurement of a single interval.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.setSampleCount(sampleCount)#
- Parameters:
sampleCount – int
Sets the number of sample points to sampleCount
. After setting the number of samples with this function, you must call create()
to instantiate the underlying OpenGL timer query objects.
The new sampleCount
must be at least 2.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.waitForIntervals()#
- Return type:
Returns a QList
containing the time intervals delimited by the calls to recordSample()
. The resulting vector will contain one fewer element as this represents the intervening intervals rather than the actual timestamp samples.
This function will block until OpenGL indicates the results are available. It is recommended to check the availability of the result prior to calling this function with isResultAvailable()
.
See also
- PySide6.QtOpenGL.QOpenGLTimeMonitor.waitForSamples()#
- Return type:
Returns a QList
containing the GPU timestamps taken with recordSample()
.
This function will block until OpenGL indicates the results are available. It is recommended to check the availability of the result prior to calling this function with isResultAvailable()
.
Note
This function only works on systems that have OpenGL >=3.3 or the ARB_timer_query extension. See QOpenGLTimerQuery
for more details.
See also