Jonathan R. Madsen

@Rombur that commit was more of an experiment on my end to see what needed to be fixed. Useful too since I didn't realize that the parse intermixed args in...

Retest this please

Retest this please

> Please split the individual changes (changes to RULE_LAUCNH_COMPILER, changes to CMake installation, using compile-time-perf) up into multiple pull requests. What? These all go together. I'll back out the changes...

I tweaked the RULE_LAUNCH_COMPILE because the compile-time-perf also uses it and the way the kokkos rule launch stuff is designed, it always needs to be last because it depends on...

@arghdos You may be interested in the [Jenkins artifacts](https://cloud.cees.ornl.gov/jenkins-ci/blue/organizations/jenkins/pykokkos-base/detail/main/2/artifacts) as part of the [pykokkos-base](https://github.com/kokkos/pykokkos-base) CI. The [CUDA-pykokkos-base-compile-time.txt](https://cloud.cees.ornl.gov/jenkins-ci/job/pykokkos-base/job/main/2/artifact/build/CUDA-pykokkos-base-compile-time.txt) and [HIP-pykokkos-base-compile-time.txt](https://cloud.cees.ornl.gov/jenkins-ci/job/pykokkos-base/job/main/2/artifact/build/HIP-pykokkos-base-compile-time.txt) are basically a head-to-head comparison of NVCC vs. HIPCC compile times...

ert_results.json can be generated from `ex_ert` with `-DTIMEMORY_BUILD_EXAMPLES=ON`

 [ert_results.json.zip](https://github.com/NERSC/timemory/files/5405511/ert_results.json.zip)

Requires changes in #89, specifically: https://github.com/NERSC/timemory/blob/88d7b915bb60c1bcf7dd0a1ab89a198ce5343a63/timemory/roofline/roofline.py#L180 ```python if self.units is not None: for i in range(len(self.data)): self.data[i] /= self.units ```

Would it help to have the L1, L2, and (if exists) L3 data cache sizes in the JSON so you can extract the ERT tests around those values?