Expose equivalent of all thrust:: algorithms in cuda:: namespace

[jrhemstad]

This issue can be closed by creating a design doc that details how we will pull various parts of thrust:: parallel algorithms into cuda:: namespace including:

• Implement std::execution execution policies in libcu++ #5590
• Design a dispatch mechanism for different parallel backends #5594
• [EPIC]: Port std equivalent parallel algorithms from thrust to cuda::std #5592
• Port non-standard thrust algorithms to cuda #5593