I'm a Computer Vision Developer.
So i'm interesting in OpenVX and now using Intel CV SDK.
But i have something problem while making custom kernel.
I 've converted my Function to Custom kernel.
But Custom kernel speed is slower than old function.
I've analyzed this reason and the time consumes while operating XOR(bool xor).
Measured times are 1.8ms/4ms(2 times slower).
Does anyone know this problem?
Please let me know this reason.
For more complete information about compiler optimizations, see our Optimization Notice.