… and the computer language benchmarks game are examples of micro benchmarks. These are easy to use, easy to measure, but far from realistic. They are nonetheless valuable tools.

My question is if anyone here has any experience with simplistic benchmarking and could tell me which things to test for in order to get a simple idea of each language's general performance?

Time to Solve TTS. Is that lowest Elapsed Time or lowest CPU Time or mean CPU Time? Optimized, Un-optimized or Goldilocks programs.

There's more than one right answer.

Answer for a few un-optimised fannkuch-redux programs —

Un-optimised single-thread #8 programs transliterated line-by-line literal style into different programming languages from the same original.

	source code	secs	mem	gz
	C gcc #8	39.34	2,343	483
	Java #8	46.66	59,994	517
	Node.js #8	58.20	63,136	452
	PHP #8	14 min	16,364	458
	Ruby yjit #8	34 min	17,691	418
	Python 3 #8	40 min	11,874	414

Or choose your answer for Rust and Julia and Haskell and … crowd sourced fannkuch-redux programs.

Answer for the fastest contributed programs —

Seven simple tasks. A couple of dozen different programming-language implementations. For each language implementation, the normalized performance of the fastest contributed program for each task.

Explore the measurements with box plot charts — medians, dispersion, skew.

Programs for binary-trees, pidits and regex-redux have not been included, because the differences in libraries and approach would be hidden in the box plot summary.

The median-line for some language implementation groups seems to fall outside the IQR box for other groups. Choose your answer for likely differences between the groups.