My goal is to use ipplib on Mac OS X Sierra .
In parallel Studio 2018 there are different sample but I not understand the better pipeline for the visualization task.
- First I tried ipp_blur_rotate_sample that come with application Cocoa for visualization. It work on command line but I have some problem to load project in Xcode .
- After I tried ipp_resize_mt and ipp_thread that work in a shell and Xcode Environment. In this example there is the source that let me think at. visualization task, but not work. This is the code:
WindowDraw draw("Intel(R) IPP Resize example", WF_FIT_TO_IMAGE);
printf("\nPress Space to cycle through stages:\n");
printf("1 - result image\n");
printf("2 - original image\n");
printf("\nClose window to exit.\n");
int iIndex = 0;
bool bRedraw = true;
if(draw.CheckKey() == KK_SPACE)
iIndex = (iIndex+1)%2;
bRedraw = true;
bRedraw = true;
if(iIndex == 0)
else if(iIndex == 1)
bRedraw = false;
} But I have no visualization or windows open or Error also if I force the if-condition at top.
I'm glad to understand what is the right way for the visualization of the result of some ipp operation on the image.
Thanks in advanced.
The functions for displaying image is defined in common folder, we uses OpenGL to read and write 2D graphics. The WindowDraw class is defined in base_renderer.h and implemented in base_renderer.cpp. You could replace to use other graphics support framework to draw img, actually Intel IPP only work for image processing, but no any support for image I/O. I am not very familiar with OpenGL drawing on cocoa view, you probably could follow some tutorial to work:
Thanks for reply.
I'll follow some tutorial for creating cocoa visualization app.
At this point I think the best starting point for visualization is ipp_blur_rotate_sample.
I'll keep you informed on this post.
You could use this sort of open source library which supports image I/O module and can be optimized with Intel IPP. There are two ways to use OpenCV with IPP, one is with special subset of Intel IPP functions for image processing and computer version which is known as IPP-ICV, the other way is with standalone Intel IPP library that should be installed separately.
You probably could follow this article for OCV3:
Enabling IPP on OpenCV
Cool..Thanks for sharing!
I think this is more linear solution for use intel c++ compiler and parallel studio tools.
However, if I had used opengl with cocoa I would have to handle the rendering pipeline on the GPU. In this case, it would not be possible to use Advisor or Vtune Amplifier to identify application bottlenecks as they do not address a Gpu pipeline survey but a memory and CPU survey.
Is the previous sentence correct ?