You would not get reliable results. There is a reason the Benchmark starts by playing the Warmup scene twice and then ignoring the measured results. You machine needs to warm up, because heat can cause throttling. If you want to compare results, other people need to run the same tests in the same sequence or its all worthless.
(And yes, there are clueless "experts" out there who managed to modify the Benchmark to run just run the last 2 tests. Despite all my efforts at integrity checks and telling everyone not to modify anything. And still post the results in the sharing thread. Completely worthless results because you don't know whether they also excluded the warmup scene or whatever

)