- 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

It used to work everywhere, but the compilers on CentOS (on penny)
are weird...

Conflicts:
	README

Conflicts:
	CMakeLists.txt
	build
	src/TNL/Assert.h
	src/TNL/Containers/SharedArray.h
	src/TNL/Containers/StaticVector.h
	src/TNL/Containers/StaticVector1D_impl.h
	src/TNL/DevicePointer.h
	src/TNL/Devices/Cuda.h
	src/TNL/Operators/diffusion/LinearDiffusion_impl.h
	src/TNL/Problems/HeatEquationProblem_impl.h
	src/TNL/SharedPointer.h
	src/TNL/Solvers/PDE/ExplicitUpdater.h
	src/TNL/Solvers/SolverMonitor.h
	src/TNL/UniquePointer.h

Conflicts:
	src/UnitTests/Containers/VectorTest.cu

(This version is implemented in Object)

Conflicts:
	examples/advection/LaxFridrichs_impl.h
	examples/inviscid-flow/1d/LaxFridrichsContinuity_impl.h
	examples/inviscid-flow/1d/LaxFridrichsEnergy_impl.h
	examples/inviscid-flow/1d/LaxFridrichsMomentum_impl.h
	examples/inviscid-flow/2d/EulerPressureGetter_impl.h
	examples/inviscid-flow/2d/EulerVelGetter_impl.h
	examples/inviscid-flow/2d/EulerVelXGetter_impl.h
	examples/inviscid-flow/2d/EulerVelYGetter_impl.h
	examples/inviscid-flow/2d/LaxFridrichsContinuity_impl .h
	examples/inviscid-flow/2d/LaxFridrichsEnergy_impl.h
	examples/inviscid-flow/2d/LaxFridrichsMomentumX_impl.h
	examples/inviscid-flow/2d/LaxFridrichsMomentumY_impl.h
	src/TNL/Meshes/GridDetails/Grid2D_impl.h
	src/TNL/Meshes/GridDetails/NeighbourGridEntitiesStorage.h
	src/TNL/Meshes/GridEntity.h
	src/TNL/Meshes/GridEntityConfig.h
	src/TNL/Operators/NeumannBoundaryConditions.h

The problem is that the __CUDA_ARCH__ macro is defined only in device
code, so it can't be used for configuring the kernel launches.