Update benchmark.js test file in order to reduce the time test-benchmark-buffer test takes to run

[fesebuv]

The following pr updates the runBenchMark function by:

1. Breaking the runBenchMark function into the getArgv for the purpose of getting the argv array.
2. When constructing the argv array, we are reducing the args array parameter. That way we make one push operation instead of two per iteration.
3. Instead of contatinating an array per iteration on line https://github.com/nodejs/node/compare/master...fesebuv:fesebuv-updates?expand=1#diff-f282ba153a1c9657bd17a47caabab8bfR29 we are pushing that line into an array and then joining that array when needed.

Run

time ./node ./test/sequential/test-benchmark-buffer

• Should see time reducing from real ~0m5.570s to real ~0m5.479s

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• documentation is changed or added
• commit message follows commit guidelines

[mscdex]

That kind of difference in results seems like it would be within the margin of error and not necessarily a consistent, measurable improvement? In general though, the last time I checked string concatenation would be faster than array pushing and joining.

[fesebuv]

🤔 perhaps. But as it was pointed out, the difference is "within the margin of error". What I am attempting to do is a little bit of code refactoring, improve readability, etc.

[apapirovski]

That difference is almost certainly within margin of error. If we want to improve performance there certainly better ways than creating an extra closure and replacing a fast for loop with a reduce that constantly reallocates arrays.

[fesebuv]

@apapirovski understood!

[fesebuv]

Based on the following comment:
#20128 (comment)

And since the updates made on this pr are not significant. I am closing this pr.
Thanks for reviewing @apapirovski @mscdex

I look forward to contribute elsewhere.