[ci] speed up conda setup for macOS and Linux jobs

[jameslamb]

Follow-up to #5668.

Proposes combining the conda create and conda install steps run in CI jobs into a single conda create step. In #5668 and #5648, we observed that doing that led to less overall time in CI spent waiting for conda solves.

benefits of this change

I observed the following timing differences in these jobs on this PR, compared to master:

job	latest master	this PR
(GHA) CUDA 11.7 wheel	12m41s	12m16s
(GHA) CUDA 10.0 source	11m17s	11m17s
(GHA) CUDA 11.7.1 pip	12m29s	12m35s
(GHA) macOS regular	16m46s	14m49s
(GHA) macOS bdist	18m33s	13m07s
(GHA) macOS sdist	13m16s	10m06s
(GHA) MPI source	16m52s	12m33s
(GHA) MPI pip	12m21s	10m29s
(GHA) MPI wheel	13m42s	13m03s
(Azure) Linux regular	14m58s	15m21s
(Azure) Linux sdist	13m29s	12m49s
(Azure) Linux bdist	16m08s	15m37s
(Azure) Linux mpi_source	10m47s	10m45s
(Azure) Linux gpu_source	16m13s	17m21s
(Azure) Linux_latest regular	16m20s	15m15s
(Azure) Linux_latest sdist	13m44s	13m52s
(Azure) Linux_latest bdist	16m38s	15m31s
(Azure) Linux_latest mpi_source	09m55s	09m53s
(Azure) Linux_latest gpu_source	16m16s	15m21s
(Azure) Linux_latest gpu_pip	16m52s	15m10s
(Azure) Linux_latest gpu_wheel	16m40s	15m41s
(Azure) QEMU_multiarch bdist	71m29s	70m41s
(Azure) macOS regular	12m08s	11m24s
(Azure) macOS sdist	10m06s	08m31s
(Azure) macOS bdist	11m12s	08m46s

this PR:

• CUDA GitHub Actions: https://github.com/microsoft/LightGBM/actions/runs/4272338290/jobs/7437402962
• macOS GitHub Actions: https://github.com/microsoft/LightGBM/actions/runs/4272338296/jobs/7437401203
• Azure: https://dev.azure.com/lightgbm-ci/lightgbm-ci/_build/results?buildId=14319&view=results

latest master build:

• CUDA GitHub Actions: https://github.com/microsoft/LightGBM/actions/runs/4273088891/jobs/7438764795
• macOS GitHub Actions: https://github.com/microsoft/LightGBM/actions/runs/4273088894/jobs/7438765167
• Azure: https://dev.azure.com/lightgbm-ci/lightgbm-ci/_build/results?buildId=14322&view=results

Notes for Reviewers

Given all the things that can change from run-to-run, it's hard to trust differences in the timings based on 1 or even a few of runs. But I'm still pretty confident that these changes make these CI jobs at least a small amount faster, without reducing their stability.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed.
To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues
including a reference to this.