On Mon, Feb 8, 2021 at 6:17 AM Heikki Linnakangas <
hlinnaka@iki.fi> wrote:
>
> I also tested the fallback implementation from the simdjson library
> (included in the patch, if you uncomment it in simdjson-glue.c):
>
> mixed | ascii
> -------+-------
> 447 | 46
> (1 row)
>
> I think we should at least try to adopt that. At a high level, it looks
> pretty similar your patch: you load the data 8 bytes at a time, check if
> there are all ASCII. If there are any non-ASCII chars, you check the
> bytes one by one, otherwise you load the next 8 bytes. Your patch should
> be able to achieve the same performance, if done right. I don't think
> the simdjson code forbids \0 bytes, so that will add a few cycles, but
> still.
That fallback is very similar to my "inline C" case upthread, and they both actually check 16 bytes at a time (the comment is wrong in the patch you shared). I can work back and show how the performance changes with each difference (just MacOS, clang 10 here):
master
mixed | ascii
-------+-------
757 | 366
v1, but using memcpy()
mixed | ascii
-------+-------
601 | 129
remove zero-byte check:
mixed | ascii
-------+-------
588 | 93
inline ascii fastpath into pg_utf8_verifystr()
mixed | ascii
-------+-------
595 | 71
use 16-byte stride
mixed | ascii
-------+-------
652 | 49
With this cpu/compiler, v1 is fastest on the mixed input all else being equal.
Maybe there's a smarter way to check for zeros in C. Or maybe be more careful about cache -- running memchr() on the whole input first might not be the best thing to do.