[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Sun, 15 Aug 2021 17:48:58 -0400
From: Rich Felker <dalias@...c.org>
To: Stefan Kanthak <stefan.kanthak@...go.de>
Cc: Szabolcs Nagy <nsz@...t70.net>, musl@...ts.openwall.com
Subject: Re: [PATCH #2] Properly simplified nextafter()
On Sun, Aug 15, 2021 at 10:52:24PM +0200, Stefan Kanthak wrote:
> Rich Felker <dalias@...c.org> wrote:
>
> > I really have better things to be doing than putting up with repeated
> > toxic interactions for the sake of a supposed miniscule improvement in
> > something nobody has identified as having any problem to begin with.
>
> Nobody as in "Niemand hat die Absicht eine Mauer zu bauen"?
> Or just nobody EXCEPT ME bothered to take a look at the code of your
> nextafter() and noticed its performance deficit (at least on AMD64
> and i386)?
I guess I missed your initial report about what software you were
running that was spending more than 1% of its execution time in
nextafter.
As I understand it, what happened was that you were reading the code
and thought "I bet I could do better". Maybe you can. But can you do
sufficiently better that it makes any practical difference, to be
worth the time spent reviewing that it's correct? Probably not.
But now, in your case, the bar is even higher. It's: Can you do
sufficiently better to be worth subjecting our good contributors to
someone who wants to insult and abuse them and demand their time doing
unproductive things? Given how fast the function already is, and how
improbable it is that it makes any difference to the run time of any
real code, the answer is most certainly no. You've raised the bar for
your own contribution so high that it can't be met.
> JFTR: your implementation is NON-COMPLIANT!
> I recommend to read the ISO C standard and follow it by the word.
Making unspecific claims lik that to waste our time trying to figure
out what you mean is not making that bar any lower.
> > If you want to engage constructively, you're welcome to. This is not
> > it.
>
> You are most obviously NOT interested in performance improvement, or
> just to stubborn: choose what you like better.
I am not interested in further subjecting the authors of our existing
code to your abuse. Please leave.
Rich
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
