[WIP] Attempt to resolve gpuCI breakage

[charlesbluca]

Not exactly sure what's causing the gpuCI failures yet, but I do notice the following popping up in our pip install logs:

ERROR: After October 2020 you may experience errors when installing or updating packages. This is because pip will change the way that it resolves dependency conflicts.

We recommend you use --use-feature=2020-resolver to test your packages with the new resolver before it becomes the default.

dask-cudf 22.2.0a0+103.g83adc4b6cf requires cupy-cuda115, which is not installed.

I am going to try out the new pip resolver to see what happens there.

[codecov-commenter]

Codecov Report

Merging #340 (ecf3dab) into main (3ac4136) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main     #340   +/-   ##
=======================================
  Coverage   95.76%   95.76%           
=======================================
  Files          65       65           
  Lines        2856     2856           
  Branches      534      534           
=======================================
  Hits         2735     2735           
  Misses         73       73           
  Partials       48       48           

---

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 3ac4136...ecf3dab. Read the comment docs.

[charlesbluca]

Resolved, looks like we had mismatched versions of cudf / libcudf due to some implicit changes in cuDF's numpy dependency from NVIDIA/cudf#9687