Update decompress_to_edgelist to handle edge types

[naimnv]

Update decompress_to_edgelist to handle edge types

[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.

[naimnv]

/okay to test

[seunghwak]

LGTM but @jnke2016 needs a decompress_to_edgelist function that takes just a graph view object and an edge_property_device_view_t object of integer types (i.e. int32_t and int64_t for now, and in the future, we may need to support floating point types as well). Do you think it is better to implement this in this PR or address this in a separate PR?

[naimnv]

LGTM but @jnke2016 needs a decompress_to_edgelist function that takes just a graph view object and an edge_property_device_view_t object of integer types (i.e. int32_t and int64_t for now, and in the future, we may need to support floating point types as well). Do you think it is better to implement this in this PR or address this in a separate PR?

He can just pass it instead of edge_id_view if edge_property_device_view_t object is of integer types (i.e. int32_t and int64_t
or pass it instead of edge_weights_view if edge_property_device_view_t object is of floating point types. And he needs it for triangle count tests, IIRC.

Otherwise we need an additional 72 explicit instantiations and I can address it in a separate PR.

[seunghwak]

Otherwise we need an additional 72 explicit instantiations and I can address it in a separate PR.

Why 72? I assume 3 (vertex_t & edge_t combinations) * 4 (store_transposed & multi_gpu) * 2 (Int32_t & int64_t) = 24. And eventually, we may remove 32 bit vertex ID and 64 bit edge ID support and may change store_transposed & multi_gpu as run-time parameters, so this will become more like 4.

[naimnv]

/okay to test

[naimnv]

/okay to test

[naimnv]

/merge