Occupancy improvement for Hash table build

[tgujar]

Description

Prototype implementation for: #15502

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[tgujar]

I think the approach of specializing the type dispatcher is very cumbersome and will lead to a lot of code replication. Currently, I have the conditional dispatch working for device_row_hasher but I am unsure if there is a better way to implement this. We could introduce a macro here to generate the code, what do you think?

[PointKernel]

/ok to test

[PointKernel]

/ok to test

[PointKernel]

@tgujar I've updated the docs to unblock CI. Have you noticed any performance regressions for other use cases? It seems that it improves the performance for mixed join but the performance drops significantly in other cases using row hasher.

[ttnghia]

I don't think decimal types are complex type. They are just a wrapper around some integer type.

[tgujar]

Equality operator for Decimal will perform scaling which uses exponentiation.

cudf/cpp/include/cudf/fixed_point/fixed_point.hpp

Line 735 in 888e9d5

CUDF_HOST_DEVICE inline bool operator==(fixed_point<Rep1, Rad1> const& lhs,

I see a reduction in register usage if I comment out decimal types in #15502. I think we can still decide on the types excluded in the branches later on

[PointKernel]

/ok to test

[PointKernel]

@tgujar Could you take a look at the failing tests?

[PointKernel]

/ok to test

[PointKernel]

/ok to test