Apply sentinel namespace to other map structures

[m3g4d1v3r]

Closes #162

Hello there!

This is my first contribution to a open source project, so please bear with me if I make a mistake or two.
I've tried to compile the last changes on my linux machine but unfortunately I've reached some errors at the benchmark compiling step.

Consolidate compiler generated dependencies of target DYNAMIC_MAP_BENCH
[ 43%] Building CUDA object benchmarks/CMakeFiles/DYNAMIC_MAP_BENCH.dir/hash_table/dynamic_map_bench.cu.o
synchronization.hpp(91): error: namespace "std" has no member "runtime_error"

Regarding the commits, I've managed to find two structures that the sentinel namespace could be applied. Please let me know if there are other structures that I've missed or if there were errors in my commits.

Best regards,
Marlon

[GPUtester]

Can one of the admins verify this patch?

[PointKernel]

@m3g4d1v3r Thanks for the contribution!

Consolidate compiler generated dependencies of target DYNAMIC_MAP_BENCH
[ 43%] Building CUDA object benchmarks/CMakeFiles/DYNAMIC_MAP_BENCH.dir/hash_table/dynamic_map_bench.cu.o
synchronization.hpp(91): error: namespace "std" has no member "runtime_error"

Which version of host compiler are you using? Include <stdexcept> in benchmarks/synchronization.hpp should fix the issue.

[ajschmidt8]

ok to test

[m3g4d1v3r]

Hello @PointKernel and @ajschmidt8 .

The inclusion of stdexcept library helped the compilation to pass the mentioned step. However, another error ocurred:

[ 55%] Building CUDA object _deps/nvbench-build/nvbench/CMakeFiles/nvbench.dir/blocking_kernel.cu.o
gcc-10: error: unrecognized command-line option ‘-Wunused-local-typedef’; did you mean ‘-Wunused-local-typedefs’?
gcc-10: error: unrecognized command-line option ‘-Wgnu’

I've used the following command to prepare the build environment:

export CC=clang && export CXX=clang++ && cmake -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ ..

[jrhemstad]

Thanks @m3g4d1v3r!

It looks like you're trying to compile with clang. We don't explicitly support this, though it is something we're looking into: #128

[m3g4d1v3r]

I see @jrhemstad!

I guess I was confused about which compiler to use, so I ended up trying to compile with clang and supposed that I was missing some configuration in the cmake step.

Following your suggestions, I've successfully compiled the project the pull request diff using gcc version 10.2.1 for Debian. As for nvcc 11+ I've used nvidia-cuda-toolkit package, also available on Debian's distro.

[PointKernel]

I have some small suggestions then it should be good to go.

[PointKernel]

Good catch

[jrhemstad]

Thanks again @m3g4d1v3r