Update benchmarks, possibly with node versions

[mattbishop]

I have found with node 12.22.1 that flatstr's performance is not that different from node now and in some cases is running slower.

Is it getting time to retire flatstr for Node 12 and higher?

unflattenedManySmallConcats10000: 164.021ms
flattenedManySmallConcats10000: 118.004ms
unflattenedSeveralLargeConcats10000: 229.177ms
flattenedSeveralLargeConcats10000: 224.262ms
unflattenedExponentialSmallConcats10000: 114.288ms
flattenedExponentialSmallConcats10000: 116.891ms
unflattenedExponentialLargeConcats10000: 3042.564ms
flattenedExponentialLargeConcats10000: 3255.192ms
unflattenedManySmallConcats10000: 138.674ms
flattenedManySmallConcats10000: 134.179ms
unflattenedSeveralLargeConcats10000: 243.686ms
flattenedSeveralLargeConcats10000: 258.039ms
unflattenedExponentialSmallConcats10000: 137.334ms
flattenedExponentialSmallConcats10000: 140.511ms
unflattenedExponentialLargeConcats10000: 3204.494ms
flattenedExponentialLargeConcats10000: 3338.138ms

[kibertoad]

@davidmarkclements Should this be mentioned in readme of the project, perhaps? I can run benchmark using https://github.com/kibertoad/photofinish in order to accumulate info per Node version.