add type trait 'remove_restrict'

[psychocoderHPC]

fix #1472

Provide a type trait to remove restrict from a type.

[j-stephan]

Is removing __restrict__ what we want to achieve here? Removing it against the user's explicit wish feels wrong to me.

[fwyzard]

Is removing __restrict__ what we want to achieve here? Removing it against the user's explicit wish feels wrong to me.

My impression is that for the CUDA and HIP backends, __restrict__ is not carried over across the call from host to device anyway. The pointers need to be declared as __restrict__ inside the kernel (on the device side).

For CPU backends the __restrict__ on the "host" side could still be meaningful (not sure about useful).

[j-stephan]

Got it, thanks!

[psychocoderHPC]

Windows CI error in all runs with debug enabled:

2021-11-26T11:32:07.5385822Z   D:\a\alpaka\alpaka\build\example\helloWorld>"C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.2\bin\nvcc.exe" -gencode=arch=compute_52,code=\"compute_52,compute_52\" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\VC\Tools\MSVC\14.29.30133\bin\HostX64\x64" -x cu   -ID:\a\alpaka\alpaka\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.2\include" -ID:\a\alpaka\alpaka\boost -I"C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.2\include"     --keep-dir x64\Debug  -maxrregcount=0  --machine 64 --compile -cudart static --extended-lambda --expt-relaxed-constexpr -lineinfo -Xcudafe=--display_error_number -Xcudafe=--diag_suppress=esa_on_defaulted_function_ignored -std=c++17 -Xcompiler="/EHsc -Zi -Ob0 /openmp" -g  -D_WINDOWS -DALPAKA_ACC_CPU_B_SEQ_T_SEQ_ENABLED -DALPAKA_ACC_CPU_B_SEQ_T_THREADS_ENABLED -DALPAKA_ACC_CPU_B_OMP2_T_SEQ_ENABLED -DALPAKA_ACC_CPU_B_SEQ_T_OMP2_ENABLED -DALPAKA_ACC_GPU_CUDA_ENABLED -DALPAKA_DEBUG=0 -DALPAKA_DEBUG_OFFLOAD_ASSUME_HOST -DALPAKA_OFFLOAD_MAX_BLOCK_SIZE=256 -DALPAKA_BLOCK_SHARED_DYN_MEMBER_ALLOC_KIB=47 -DALPAKA_CI -D"CMAKE_INTDIR=\"Debug\"" -D"CMAKE_INTDIR=\"Debug\"" -D_MBCS -Xcompiler "/EHsc /W1 /nologo /Od /FdhelloWorld.dir\Debug\vc142.pdb /FS /Zi /RTC1 /MDd /GR" -o helloWorld.dir\Debug\helloWorld.obj "D:\a\alpaka\alpaka\example\helloWorld\src\helloWorld.cpp" 
2021-11-26T11:32:13.7723136Z   helloWorld.cpp
2021-11-26T11:32:14.9954951Z C:\Users\runneradmin\AppData\Local\Temp\tmpxft_000007fc_00000000-7_helloWorld.cudafe1.stub.c(27): 
error C2912: explicit specialization 'void alpaka::uniform_cuda_hip::detail::
__wrapper__device_stub_uniformCudaHipKernel<alpaka::AccGpuCudaRt<std::integral_constant<unsigned __int64,3>,unsigned __int64>,std::integral_constant<unsigned __int64,3>,unsigned __int64,HelloWorldKernel>(const _ZN6alpaka3VecISt17integral_constantIyLy3EEyEE &,const HelloWorldKernel &)' 
is not a specialization of a function template 
[D:\a\alpaka\alpaka\build\example\helloWorld\helloWorld.vcxproj]
2021-11-26T11:32:15.0268017Z C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\MSBuild\Microsoft\VC\v160\BuildCustomizations\CUDA 11.2.targets(785,9): error MSB3721: The command ""C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.2\bin\nvcc.exe" 

[psychocoderHPC]

Switched to draft. I changed some lines to pass the windows CI. I will run Monday some more tests if all CI jobs passed.

[psychocoderHPC]

This PR is now passing the CI and can be reviewed.