Re: Performance improvements for src/port/snprintf.c

On Thu, Sep 27, 2018 at 1:18 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> On 2018-09-26 17:57:05 -0700, Andres Freund wrote:
> > snprintf time = 1324.87 ms total, 0.000264975 ms per iteration
> > pg time = 1434.57 ms total, 0.000286915 ms per iteration
> > stbsp time = 552.14 ms total, 0.000110428 ms per iteration
>
> Reading around the interwebz lead me to look at ryu
>
> https://dl.acm.org/citation.cfm?id=3192369
> https://github.com/ulfjack/ryu/tree/46f4c5572121a6f1428749fe3e24132c3626c946
>
> That's an algorithm that always generates the minimally sized
> roundtrip-safe string output for a floating point number. That makes it
> insuitable for the innards of printf, but it very well could be
> interesting for e.g. float8out, especially when we currently specify a
> "too high" precision to guarantee round-trip safeity.

Wow. While all the algorithms have that round trip goal, they keep
doing it faster. I was once interested in their speed for a work
problem, and looked into the 30 year old dragon4 and 8 year old grisu3
algorithms. It's amazing to me that we have a new algorithm in 2018
for this ancient problem, and it claims to be 3 times faster than the
competition. (Hah, I see that "ryū" is Japanese for dragon. "Grisù"
is a dragon from an Italian TV series.)

--
Thomas Munro
http://www.enterprisedb.com