[code health] consolidate away from separate docker/slurm benchmark scripts

[cquil11]

There is no real reason to have different benchmarks/MODEL_PRECISION_GPU_FRAMEWORK_[slurm|docker].sh. The code inside these files runs inside a container, either on a SLURM or Docker machine -- either way the code is the same (with the exception of a few trivial differences like printing SLURM Job ID).

In order to improve code health and reduce the number of files that need to be maintained. We can consolidate these.