Running versus designing experiments

I hope that the person i aim to read came through this article, however this is an interesting point that fresh & moderate experience engineer might experience at least once through their time. i was quoted in one of the monthly report of another highly respected personnel  at my work place.

As a little background on the quote, the personnel  was coming to several of us during lunch time with his problem: :running a Perl scripts which took 15 minutes at Intel core gen3 vs. 13 minute at Intel dual core vs. 10 minute on a Xeon machine.

As first statement that I asked is are those running on a same OS?

The answer is no. so I quickly came with a hypothesis that that running on Linux is a lot faster due to less overhead on running interpreter such as Perl provided sufficient RAM was available and no swap file was generated.

However my advice during the time was to run using a Cygwin (a Linux environment created under window)

And without too much thought, the experiments take place & the result was not aligning with the hypothesis that I have.

Revising my thought over tea break in Intel (which I really like to do it by myself to reflect to what I’ve done on that particular day) I realize that the experiments introduce more noises where Cygwin overhead was based on windows command prompt.

Best experiments is by creating an identical VM with different OS running the same script

Design of experiment (DOE) matrix

                                                |Windows  | Linux

512MB RAM, Single Core:        |1500sec+  | 2000sec+

2GB RAM Single Core              |1300sec + | 800sec+

After forwarding the result to the personnel ,

I was quite confident that the DOE was a conclusive to support his work & effort.

However today, the personnel monthly report mail quoted my idea, effort as

"Sought help to run some experiments"

I was really disappointed since coming out with the controlled parameters experiments is way more that running the experiment. It was not part of my Job role & the least that I will get by helping is the proper credit (via email is fine). However the credit quoted to me was translated to some guy spent his time doing the data collection.

kind of sad, but that is what happened  

Please mind your communication/report (general advice) since you don't want 

the person that helped you turn you away on the next help you seek.