used BufferPool mechanism instead

the algorithm is not stable yet

use correct types for integral/sum outputs

* improve `CUDA_TARGET_CPU_ARCH` cache initialization,
    allow to override initial value from calling script;
* add `CUDA_TARGET_OS_VARIANT` option to select OS variant;
* add `CUDA_TARGET_TRIPLET` option to select target triplet from
    `${CUDA_TOOLKIT_ROOT_DIR}/targets` folder;
* remove `CUDA_TOOLKIT_TARGET_DIR` option, now it is calculated from
    `CUDA_TARGET_TRIPLET`, the old approach still can be used for compatibility;
* for CUDA 6.5 and newer try to locate static libraries too, because
   in 6.5 toolkit for ARM cross compilation only static libraries are included.

add swap instruction for allocator field

use CV_CUDA_TEST_MAIN for opencv_test_core to initialize CUDA
device information

it allows to use cudaHostAlloc methods for cv::Mat objects

* it is the only place, where it is used
* no need to make this function public

move main CUDA group to modules/core/cuda.hpp

add a note to use new cudev module as a replacement

it is internal class, no need to export it

The deinitialization of BufferPool internal objects is controled by global
object, but it depends on other global objects, which leads to errors
caused by undefined deinitialization order of global objects.

I merge global objects initialization into single class, which performs
initialization and deinitialization in correct order.