This reverts commit dbce9069.

it fails because driver terminates CUDA kernels after time out

do not calculate sum of error in every round of iteration;
instead the error will be summed every 2nd times or more, 
if the previous sum of error is too far away from threshold.

Looks like we don't actually use them; when creating precompiled headers
with Visual C++, a dummy .cpp is created at build time.

Turns out, you can't use preprocessor directives inside macro arguments.
Who'd have thought?

Now it allows choosing between arbitrary implementation variants.

+ interpolateFrames and createOpticalFlowNeedleMap

Also cv::, cv::gpu:: and cv::ocl:: namespace prefixes can be safely omitted
inside CV_ENUM and CV_FLAGS

* disabled warnings from thrust
* fixed warnings from ts_gtest.h
* possibly fixed superres compilation in Debug mode on Windows

* added DenseOpticalFlow interface
* moved OpticalFlowDual_TVL1 to src folder