'High Performance' Slow Relative to 'Balanced'

[#1]

I have a script that I've developed that takes about 50 seconds to run.  But when I move it onto another laptop it takes about 500 seconds to run (same version of JMP, different machine specs but arguably the slower machine is more powerful - faster processor, more RAM).  Strangely, if I change the battery setting from High Performance to Balanced, the execution time goes down from 500 seconds to 200 seconds.

75% of the execution time is dominated by row selections, the rest by calculations.  I've tried to create a test script that mimics some of the behaviour - attached.  I've not fully succeeded in reproducing the performance difference between the 2 machines but I still see the strange dependency on battery setting:

High Performance: 89 seconds

Balanced: 38 seconds

If you want to try the script, upon completion it produces a table with a script 'view log'.  This displays the execution time.

Any thoughts why the performance degrades with the high performance option?

[#2]

Both machines are Dell XPS running Windows 10.

[#3]

Any chance the slower laptop runs hotter?  Perhaps the cooling coils cannot keep up with 100% CPU load all the time. They could be dusty or just not big enough.

[#4]

PS: With three runs on my laptop the power setting made no difference:

Balanced: 42.3
High Performance: 39.7

Prob > |t| = 0.25

[#5]

Not an answer, but has related information: https://support.microsoft.com/en-us/help/2207548/slow-performance-on-windows-server-when-using-the-b...

[#6]

David,

Many years ago, I had a DEC laptop, that had battery issues.  It's visible symptom was an intermitent mouse pointer.  DEC investigated and it was discovered that the DEC design had the battery and the mouse using the same Interrupt.  Something like this might be going on with your issue.  

[#7 · Accepted solution]

Thanks for the responses. They helped me take a look at this from a different perspective - I was thinking more about my code than my hardware configuration.  It turned out that my power management settings when running on battery were somehow switched - balanaced was using full CPU power whereas High Performance had CPU usage throttled to 50%.  Having fixed that performance is now consistent across the settings.