Looks like it’s just the IronPython implementation that’s much slower. The standard CPython implementation is on par with @laurent_delrieu’s C# script. But then again the CPython component comes with a potential input cost on large data sets:
250522_StandardDeviation_00.gh (15.7 KB)