cudev is used for new device layer

it uses HAVE_CUDA, HAVE_NVCUVID, etc. macroses

used new CV_PERF_TEST_MAIN_WITH_IMPLS mechanism

* converted it to Algorithm
* old API still can be used for source compatibility (marked as deprecated)

* converted it to Algorithm, because implementation uses inner buffers and
  requires preprocessing step
* new implementation splits preprocessing and transform,
  what is more effecient
* old API still can be used for source compatibility (marked as deprecated)

old version isn't compiled with CUDA 5.5
new version doesn't depend on functional.hpp

This reverts commit fba72cb6, reversing
changes made to 02131ffb.