[dippersauce]

The unified architecture, combined with the speed of the modules (4266 MHz LPDDR4X), combined with the close physical proximity to the CPU cores likely all contribute to the M1's memory performance.

In Apple's own words:

“M1 also features our unified memory architecture, or UMA. M1 unifies its high‑bandwidth, low‑latency memory into a single pool within a custom package. As a result, all of the technologies in the SoC can access the same data without copying it between multiple pools of memory. This dramatically improves performance and power efficiency." [0]

[0] https://www.apple.com/mac/m1/

[kristianp]

Probably the speed of the modules, the mbw[1] tool does copies to do its measurement, so the close proximity and reduced copies in the OS wouldn't be factors.

[1] https://github.com/raas/mbw