> This is problematic because not only do we get an extra read

I don't understand why you're speculating on this. The author provided the assembly that only has 1 read.

FWIW, I just tested this on my machine. I get the following results:

Intel Core i7-7800 3.5Ghz / Ubuntu 20 / g++ 9.3.0 Rolled loop : 2.08 Gflops | 16.7 GB/s bandwidth 8-unrolled loop : 2.48 Gflops | 19.8 GB/s bandwidth

Here's the unrolled assembly (8 iters), with -O3:

    .L3:
            addl    (%rdi), %eax
            addq    $32, %rdi
            movl    %eax, -32(%rdi)
            addl    -28(%rdi), %eax
            movl    %eax, -28(%rdi)
            addl    -24(%rdi), %eax
            movl    %eax, -24(%rdi)
            addl    -20(%rdi), %eax
            movl    %eax, -20(%rdi)
            addl    -16(%rdi), %eax
            movl    %eax, -16(%rdi)
            addl    -12(%rdi), %eax
            movl    %eax, -12(%rdi)
            addl    -8(%rdi), %eax
            movl    %eax, -8(%rdi)
            addl    -4(%rdi), %eax
            movl    %eax, -4(%rdi)
            cmpq    %rdi, %rdx
            jne     .L3