Need to specify CMAKE_CUDA_COMPILER to clang.

Moving tag_dispatch to separate module

Seperate header and source to enable datapar tests.

Rename HPX_WITH_CUDA_COMPUTE with HPX_WITH_COMPUTE_CUDA

Improve performance of Spinlocks

in order to match the name of the module.

Fix find_end parallel overload

Adapt datapar copy

Make performance plot show even if similar performance

Post 1.7.0 version bump

Uses std::findend to compare the results
Removes unused variable count

Replaces typedef reference with auto.
Removes unused curr Iterator variable.
Uses mid and mid2 t...

Change the print-bind output to be more precise

Properly support ASAN with MSVC

Use more verbose names in sender algorithms

Final release notes for 1.7.0

Add back explicit -pthread flags when available

Adapt uninitialized_fill and uninitialized_fill_n to C++ 20

Adapt uninitialized_copy and uninitialized_copy_n to C++ 20

Update CUDA polling logging to be more verbose

Fix when_all_sender connect for references

Rename the new P0443/P2300 executor to thread_pool_scheduler

This adds support for systems that assume all types are bitwise serializable by default

Delete datapar/foreach_tests.hpp

removes foreach_tests.hpp which was accidently copied into datapar unit tests in previous commits.

- flyby: disable bitwise serialization if endianess differs

Stored senders should not be moved when using the lvalue reference
overload of connect.

Add deprecation warnings for v1.8

Adding execution::experimental::bulk algorithm

- this also allows for certain types to be normally serialized, still

Revert "Adding the missing defined for HPX_HAVE_DEPRECATION_WARNINGS"

Adapt uninitialized_value_construct and uninitialized_value_construct_n to C++ 20

Adapt uninitialized_default_construct and uninitialized_default_construct_n to C++ 20

Disable Vc in final docker image created in CI

Adding more concept checks for sender/receiver algorithms

Adapt uninitialized_move and uninitialized_move_n to C++ 20