mtspr 9, 3             # move r3 to CTR
setvl 3,0,8,0,1,1      # MVL=8, VL=r3=MIN(MVL,CTR)
# load VL bytes (update r4 addr) at width=8 (dw=8)
addi 6, 0, 0           # set all 64-bits of r6=0
sv.lbzu/pi/dw=8 *6, 1(4)
gbbd 8,6 # gbbd performs the transpose
# now bits are turned around, popcnt and sum them
setvl 0,0,8,0,1,1       # set MVL=VL=8
sv.popcntd/sw=8 *24,*8  # do (transposed) popcnt
sv.add *16,*16,*24      # accumulate in results
# branch back if CTR non-zero works even when VL=8
sv.bc/all 16, *0, -0x28 # reduces CTR by VL