- new reduce implementation (with kepler optimizations)
- saturate_cast via asm command
- video SIMD instructions in element operations
- float arithmetics instead of double
- new deviceSupports function

fixed several problems with CUDA 5.0
* gpu::LUT, uses device memory instead of host memory
* gpu::multiply, round mod for CV_8U depth
........

* gpu::LUT, uses device memory instead of host memory
* gpu::multiply, round mod for CV_8U depth

added assertion on double types for old devices

fixed some bugs in gpu module

removed void integral(const GpuMat& src, GpuMat& sum, GpuMat& sqsum, Stream& stream) - it fails with NPP_NOT_IMPLEMENTED error
updated docs, accuracy and performance tests

fixed gpu module error reporting
added asynchronous version of some functions

2) NVidia tests for GPU
3) FD sample that uses NVidia's interface.

Only for VS2008 now.
Sample for it.
new NPP_staging for VS2008 only

added mask support to SURF_GPU.

refactored gpu module, added vec math operators for uint, added support of 2 channel images into gpu::sum (removed support of double)

changed buffer type in linear filters to float.
added support of 1 channel image to linear filters.
added support of BORDER_REFLECT101, BORDER_REPLICATE and BORDER_CONSTANT border type to gpu linear filters.
minor fix in tests.
update comments in gpu.hpp.

refactored transpose_gpu, made it non template function.

fixed compile error in transform.

added better threads configuration estimator for the minMax, minMaxLoc, countNonZero functions in gpu module

removed double support (gpu: minMax, minMaxLoc, countNonZero) for CC which doesn't have native double support