Adjust gbench include_dir for CI permissions

[wphicks]

Explicitly install gbench headers in include/benchmark, since dockerized CI runs do not have permission to access include directly

[wphicks]

@dantegd I think this should do it.

[robertmaynard]

Can somebody bring me up to speed on why does CUML need to install gbench at all?

[dantegd]

@robertmaynard I don't think it needs to be installed as can be seen in the PR I opened before #4049, we thought that gtest was also being installed so this was a more consistent approach, but now that I see the logs I think my memory betrayed me and we're not installing gtest so perhaps not installing gbench would be the way to go

[robertmaynard]

and we're not installing gtest so perhaps not installing gbench would be the way to go

I vote for not installing it, since it 'should be' an implementation detail of the benchmarks which CUML doesn't install

[wphicks]

I don't have strong feelings on this either way, but if we're not installing gtest, there's certainly no need to install gbench. @dantegd any objection to me closing this out in favor of #4049?

[codecov-commenter]

Codecov Report

❗ No coverage uploaded for pull request base (branch-21.08@f654048). Click here to learn what that means.
The diff coverage is n/a.

@@               Coverage Diff               @@
##             branch-21.08    #4050   +/-   ##
===============================================
  Coverage                ?   85.59%           
===============================================
  Files                   ?      230           
  Lines                   ?    18221           
  Branches                ?        0           
===============================================
  Hits                    ?    15596           
  Misses                  ?     2625           
  Partials                ?        0           

Flag	Coverage Δ
dask	48.14% <0.00%> (?)
non-dask	77.92% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out 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 f654048...58ef833. Read the comment docs.

[wphicks]

If we do decide to go this route, we need to merge upstream changes back in. Not doing it yet to avoid an unnecessary CI run until we decide for sure whether or not we're using this PR.