Use mambabuild to build conda packages

[jjacobelli]

Use mambabuild to build conda packages. This should speed up the builds and help to debug conda conflict issues

[jjacobelli]

rerun tests

[jjacobelli]

rerun tests

[bdice]

@Ethyling I have closed my previous PR #9953 in favor of this PR. The previous PR has some helpful discussion / questions that I believe are resolved by the use of conda compilers in #10275.

[jjacobelli]

@Ethyling I have closed my previous PR #9953 in favor of this PR. The previous PR has some helpful discussion / questions that I believe are resolved by the use of conda compilers in #10275.

Yes, I think using conda compilers solved the issues

[jjacobelli]

rerun tests

[jjacobelli]

rerun tests

[codecov]

Codecov Report

Merging #10911 (c6260eb) into branch-22.06 (54789ee) will increase coverage by 0.03%.
The diff coverage is 93.75%.

@@               Coverage Diff                @@
##           branch-22.06   #10911      +/-   ##
================================================
+ Coverage         86.30%   86.33%   +0.03%     
================================================
  Files               144      144              
  Lines             22665    22681      +16     
================================================
+ Hits              19560    19581      +21     
+ Misses             3105     3100       -5     

Impacted Files	Coverage Δ
python/cudf/cudf/core/indexed_frame.py	91.70% <ø> (ø)
python/cudf/cudf/utils/ioutils.py	79.47% <87.50%> (-0.13%)	⬇️
python/cudf/cudf/core/dtypes.py	97.01% <92.30%> (-0.19%)	⬇️
python/cudf/cudf/core/dataframe.py	93.78% <100.00%> (+0.04%)	⬆️
python/cudf/cudf/core/frame.py	93.43% <100.00%> (+0.01%)	⬆️
python/cudf/cudf/io/avro.py	78.57% <100.00%> (ø)
python/cudf/cudf/io/csv.py	91.80% <100.00%> (ø)
python/cudf/cudf/io/json.py	97.56% <100.00%> (ø)
python/cudf/cudf/io/orc.py	92.77% <100.00%> (ø)
python/cudf/cudf/io/parquet.py	90.83% <100.00%> (ø)
... and 7 more

---

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 fd5724f...c6260eb. Read the comment docs.

[bdice]

Great work @Ethyling! This looks like a big win from the preliminary CI data. I compared the build times of the three most recent merged PRs, all of which were close to the numbers in the "Before" column. The "After" column is from just this PR. edit: I also verified that these "after" times are close to the three most recent CI runs since this PR was merged. It's real! The CI time savings here are on the order of 33%! With the current workflow / parallel builds, that's about 45 minutes faster start-to-finish, and about 144 minutes less CI time burned.

CI Job	Before (#10981)	After (this PR)
python-arm64/3.8/cuda/11.5	32 min	23 min
python-arm64/3.9/cuda/11.5	33 min	22 min
python/3.8/cuda/11.5	45 min	28 min
python/3.9/cuda/11.5	43 min	28 min
gpu/cuda/11.0/driver-450/python/3.8/centos7	1 hr 32 min	1 hr 2 min
gpu/cuda/11.2/driver-495/python/3.9/ubuntu18.04	1 hr 31 min	1 hr 1 min
gpu/cuda/11.5/driver-495/python/3.9/ubuntu20.04	1 hr 29 min	57 min