Comment 13 for bug 1951032

Some more results:

= Raspberry Pi 3B 1GB =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 48.24 | 46.19 | -4.26%
    65536 | 85.99 | 79.96 | -7.02%
   131072 | 154.00 | 139.68 | -9.30%
   262144 | 178.72 | 164.12 | -8.17%
   524288 | 163.56 | 156.55 | -4.28%
  1048576 | 246.15 | 234.32 | -4.81%

= Raspberry Pi 3A+ 512MB =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 57.11 | 54.22 | -5.06%
    65536 | 101.16 | 94.53 | -6.56%
   131072 | 186.94 | 168.37 | -9.94%
   262144 | 200.16 | 181.37 | -9.39%
   524288 | 175.91 | 168.93 | -3.97%
  1048576 | 261.19 | 250.62 | -4.04%

= Raspberry Pi Zero 2 =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 40.58 | 38.75 | -4.51%
    65536 | 72.51 | 67.57 | -6.81%
   131072 | 132.02 | 121.20 | -8.20%
   262144 | 165.26 | 149.13 | -9.76%
   524288 | 160.46 | 153.15 | -4.55%
  1048576 | 241.92 | 230.87 | -4.57%

Worth noting that the Pi 4 uses the 2711 SoC, while these (the 3B, 3A+, and Zero 2) all use the older 2837 SoC. In other words, while the new memcpy seems "okay" on the 2711, it's got "some" performance regression on the 2837.