Use QtQuick with cuda 8.0 cause the Application blinking when resize

tham

OS : win10 64bits
laptop : Y410P

I am developing an object recognition app with dlib cnn module, if I do not link to cuda, everything works fine, but if I link to cuda, the screen will blink(turn to black) when I resize the application window.

Following are my .pro file. I do not register any type or use any class which manipulate cuda in qml parts, what I did is compile and link to the object file(those cpp files) which leverage cuda and cuDNN.

QT += qml quick

CONFIG += c++11

INCLUDEPATH += ../

DEFINES += DLIB_USE_CUDA
DEFINES += DLIB_JPEG_SUPPORT
DEFINES += DLIB_PNG_SUPPORT

CUDA_LIB = C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v8.0/lib/x64

LIBS += "$${CUDA_LIB}/cublas.lib"
LIBS += "$${CUDA_LIB}/cublas_device.lib"
LIBS += "$${CUDA_LIB}/cuda.lib"
LIBS += "$${CUDA_LIB}/cudadevrt.lib"
LIBS += "$${CUDA_LIB}/cudart.lib"
LIBS += "$${CUDA_LIB}/cudnn.lib"
LIBS += "$${CUDA_LIB}/cufft.lib"
LIBS += "$${CUDA_LIB}/curand.lib"

include(../../pri/dlibs.pri)

SOURCES += main.cpp \
    ../logics/smart_eye.cpp \
    ../logics/smoke_eye.cpp

RESOURCES += qml.qrc \
    icons.qrc \
    train_model.qrc

# Additional import path used to resolve QML modules in Qt Creator's code model
QML_IMPORT_PATH =

# Default rules for deployment.
include(deployment.pri)

HEADERS += \
    ../logics/smart_eye.hpp \
    ../logics/smoke_eye.hpp \
    ../logics/smoke_network.hpp

How could I fix this issue?Thanks