[FEA]: Design API for group-based cooperative algorithms

[gevtushenko]

Is this a duplicate?

• I confirmed there appear to be no duplicate issues for this request and that I agree to the Code of Conduct

Area

CUDA Experimental (cudax)

Is your feature request related to a problem? Please describe.

We don't have group-based cooperative algorithms.

Describe the solution you'd like

This issue can be closed with a design of the group-based cooperative algorithms that addresses the following questions:

• how to query and pass temporary storage
• how to specify temporary storage memory space (global, shared, etc)
• how to express where output is valid
• how to pass tuning knobs

Describe alternatives you've considered

No response

Additional context

No response