Dgl graph store

[BradReesWork]

new GraphStore for DGL intergration
Code will need to be added to DGL

[codecov-commenter]

Codecov Report

❗ No coverage uploaded for pull request base (branch-22.04@afeebee). Click here to learn what that means.
The diff coverage is n/a.

@@               Coverage Diff               @@
##             branch-22.04    #2046   +/-   ##
===============================================
  Coverage                ?   73.27%           
===============================================
  Files                   ?      154           
  Lines                   ?     9947           
  Branches                ?        0           
===============================================
  Hits                    ?     7289           
  Misses                  ?     2658           
  Partials                ?        0           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update afeebee...c786ae1. Read the comment docs.

[BradReesWork]

@gpucibot merge

[wangxiaoyunNV]

Hi Brad,
When I review the graphStorage class, I notice here is no output_device parameter in the sampling functions. But the DGL graphstorage sampling APIs have output_device parameter. Andrei's code has that as well. Could wee add that parameter to the samplings functions? or have another function called to_device in the GraphStorage class?
https://github.com/dmlc/dgl/blob/701b4fccc2eed979ae3db801fabb6bf7bc03940c/examples/pytorch/__temporary__/dglnew/graph/graph.py

Thank you,
Xiaoyun

[wangxiaoyunNV]

I have a concern about CuGraphStore class, it should have a to_device function or parameter in the sampling functions.

Thank you!

[wangxiaoyunNV]

Can we also have a function like async_fetch(self, ids, device): in the cuGraphStore class? so we can send the sampled graphs to different GPUs.
Thank you!