Yunsong Wang

Sorry for the misclicking. Just wanted to check some of the commits...

> if the search space has no/little duplicates, using static_multimap + pair_contains is significantly slower This is expected since multimap is optimized to handle duplicates. i.e. if few duplicates are...

Just to make sure: does `time old` correspond to the `static_map` implementation? I assume `data X10` refers to multiplicity == 10, right?

> this is another case where making progress on the unification of different cuco map types would be beneficial. Time to kick off the[ grand refactoring ](https://github.com/NVIDIA/cuCollections/issues/110)then. `atomic_ref` is more...

> > Time to kick off the[ grand refactoring ](https://github.com/NVIDIA/cuCollections/issues/110)then. `atomic_ref` is more or less ready so I think we are good to go. > > Because you mentioned `atomic_ref`,...

@bwyogatama We should move the `function.cu` file to a place like `cudf/cpp/src/udf/kernels.cu` (just the first thing came into my mind, feel free to suggest the proper place) to make this...

add to allowlist

@gpucibot merge

The issue is still relevant since the error disappeared in nvcc/11.5.0 (see https://godbolt.org/z/P9h9jnsen) but comes back again with newer versions (`V11.5.50` and `V11.5.119`): ```bash (base) yunsongw@yunsongw-dt:~/Work/toto-cuda/test$ make nvcc -std=c++17 -arch=sm_75...

To clarify, I was testing the libcu++ version alongside `11.5.0` and `11.5.119`.