Re: [PATCH] Hex-coding optimizations using SVE on ARM.

> Thanks.  I noticed that this stuff is simple enough that we can use
> port/simd.h (with a few added functions).  This is especially nice because
> it takes care of x86, too.  The performance gains look similar to what you
> reported for v6:

This looks good, much cleaner.
One possible improvement would be to use a vectorized table lookup instead of compare and add. I compared v6 and v7 Neon versions, and v6 is always faster.
I’m not sure if SSE2 has a table lookup similar to Neon.

arm - m7g.4xlarge

  buf  | v6-Neon| v7-Neon| % diff

-------+--------+--------+--------

    64 |   6.16 |   8.57 |  28.07

   128 |  11.37 |  15.77 |  27.87

   256 |  18.54 |  30.28 |  38.77

   512 |  33.98 |  62.15 |  45.33

  1024 |  64.46 | 117.55 |  45.16

  2048 | 124.28 | 254.86 |  51.24

  4096 | 243.47 | 509.23 |  52.19

  8192 | 487.34 | 953.81 |  48.91

-----
Chiranmoy