Math.h: disable scalar functions for array/class types to avoid possible ambiguity with vector expressions

Moved (most of) static methods from TNL::Devices::Cuda as free functions into separate namespace TNL::Cuda

The class TNL::Devices::Cuda was too bloated, breaking the Single
Responsibility Principle. It should be used only for template
specializations and other things common to all devices.

The functions in MemoryHelpers.h are deprecated, smart pointers should
be used instead.

The functions in LaunchHelpers.h are temporary, more refactoring is
needed with respect to execution policies and custom launch parameters.

Also, templates were reordered so that related functions are next to
each other.

Nina Džugasová authored 6 years ago and Tomáš Oberhuber committed 6 years ago

Added ConfigDescription documentation and added small changes in Object, Math and ParameterContainer.

Added ConfigDescription documentation and added small changes in Object, Math and ParameterContainer.

The std::isnan method can be safely used in host code (if the compiler
is not moron). There is no device code which uses isnan so the
TNL::isNan function is useless.

This reverts commit 0eff2f63.

TODO: check later if it is really necessary

Implementation of TimeIndependentPDESolver.

- Cleaned up unnecessary specializations of TNL::max and TNL::min
- Implemented result type promotion to fix compilation with nvcc
- Added TNL::sqrt, which will be necessary for MIC

Implementace Array Operations pro MIC, vedoucí k zpřehlednění kódu -> kód použitý pro srovnávací test MIC-krakow, MIC-sharedpoiters, MIC-ArrayOperations