@dan given that with sufficiently bad board design (interference from the power ciruitry into the audio circuitry) you can hear code behavior such as CPU wakeups, i would be curious to see someone take the documented situation, discard any and all explanations given in that forum (which are trash), and investigate if there's a measurable difference

different memcpy implementations might activate SIMD units in the CPU or not, which may be measurable.

new vs. malloc is definitely nonsense though