- 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 printCudaDeviceInfo for new CC
- fixed some compilation errors and warnings
- removed unset command from CMake script
- removed unused std imports

Introduced new Java API
void org.opencv.core.Core.setErrorVerbosity(boolean verbose)
used to suppress output to stderr from OpenCV's asserts

* use OPENCV_TEMP_PATH environment variable on all platforms
* fix cleanup after OpenCV tests on Windows
* add --list flag to output names of all tests found
* do not override user-passed --perf_min_samples and --perf_force_samples
  options by --check flag
* fix complier checks inside run.py

There was missing windows.h include in OpenGL interop code.

* typo in Algorithm::addParam for 'short' data type which results in segmentation fault on ARM
  on attempt to set type property for FAST using setter inherited from Algorithm
* added short type support to read/write methods
* improved some error messages

made it to give more verbose errors when arguments of wrong types are passed, added setters with types (e.g. setInt, etc)

After setting depth to CV_32F for non double types this produces difference in results. I've attached test that reproduce this issue.The issue was found during anali broken GPU tests  for multiply/device functions.

This fixes build without TBB and CUDA. Issue #2445

This partly reverts commit 6ca61827.

* Refactor auto-detection of parallel frameworks
* Implement cv:getNumThreads, cv::setNumThreads and cv::getThreadNum for all supported frameworks
* From now cv::setNumThreads(0) can be used to turn off parallelisation

added recommended number of stripes to parallel_for_, modified some of the functions to use larger stripes (for better performance)