From: Joseph Myers <joseph@codesourcery.com>
To: Patrick McGehearty <patrick.mcgehearty@oracle.com>
Cc: <libc-alpha@sourceware.org>
Subject: Re: [PATCH] Improves __ieee754_exp() performance by greater than 5x on sparc/x86.
Date: Thu, 14 Dec 2017 01:28:00 -0000 [thread overview]
Message-ID: <alpine.DEB.2.20.1712140121380.5542@digraph.polyomino.org.uk> (raw)
In-Reply-To: <1512774211-43942-1-git-send-email-patrick.mcgehearty@oracle.com>
On Fri, 8 Dec 2017, Patrick McGehearty wrote:
> Revised constants t2, t3, t4, t5 to better match values of 1/n factorial.
To expand on the logic for such a change:
If the values were previously not 1/n! presumably they were coefficients
in some form of minimax approximation minimising the maximum error
(however measured) in the interval used in the original implementation.
The maximum error from just using 1/n! would be at the endpoints of the
interval (whereas a minimax approximation using an nth degree polynomial
would have equal maximum errors with alternating signs at n+2 points -
increasing some errors closer to 0 to decrease those at the endpoints).
You've changed the code to use a narrower interval. Thus, the original
minimax approximation is no longer optimal for the new interval, and it's
quite plausible that the maximum error from using 1/n! is smaller when you
restrict to the new interval.
This patch version is OK.
--
Joseph S. Myers
joseph@codesourcery.com
next prev parent reply other threads:[~2017-12-14 1:28 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-12-08 23:08 Patrick McGehearty
2017-12-11 8:14 ` Siddhesh Poyarekar
2017-12-11 17:04 ` Patrick McGehearty
2017-12-11 17:53 ` Siddhesh Poyarekar
2017-12-14 1:28 ` Joseph Myers [this message]
2017-12-18 20:11 ` Patrick McGehearty
-- strict thread matches above, loose matches on Subject: below --
2017-12-29 23:42 Patrick McGehearty
2018-01-01 1:36 ` Joseph Myers
2018-01-01 16:31 ` Patrick McGehearty
2018-01-01 16:41 ` Joseph Myers
2017-12-04 21:53 Patrick McGehearty
2017-12-05 23:20 ` Joseph Myers
2017-12-01 0:51 Patrick McGehearty
2017-12-01 0:56 ` Joseph Myers
2017-11-07 4:25 Patrick McGehearty
2017-11-16 17:52 ` Patrick McGehearty
2017-11-16 18:27 ` Carlos O'Donell
2017-11-16 18:31 ` Joseph Myers
2017-11-23 21:19 ` Joseph Myers
2017-12-01 0:47 ` Patrick McGehearty
2017-10-26 22:53 Patrick McGehearty
2017-11-01 0:26 ` Joseph Myers
2017-10-26 16:44 Patrick McGehearty
2017-10-26 17:20 ` Joseph Myers
2017-10-26 17:25 ` Joseph Myers
2017-10-26 18:30 ` Patrick McGehearty
2017-10-26 19:44 ` Joseph Myers
2017-10-20 13:38 Wilco Dijkstra
2017-10-20 14:58 ` Patrick McGehearty
2017-10-16 16:56 Patrick McGehearty
2017-10-18 17:22 ` Joseph Myers
2017-10-18 23:22 ` Joseph Myers
2017-10-19 22:31 ` Patrick McGehearty
2017-10-19 22:48 ` Joseph Myers
2017-10-20 15:04 ` Patrick McGehearty
2017-10-21 5:23 ` Patrick McGehearty
2017-10-23 12:47 ` Joseph Myers
2017-10-23 19:58 ` Patrick McGehearty
2017-10-23 21:31 ` Joseph Myers
2017-10-20 11:41 ` Szabolcs Nagy
2017-10-20 14:56 ` Patrick McGehearty
2017-10-20 16:10 ` Joseph Myers
2017-10-23 12:25 ` Siddhesh Poyarekar
2017-10-23 15:58 ` Joseph Myers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=alpine.DEB.2.20.1712140121380.5542@digraph.polyomino.org.uk \
--to=joseph@codesourcery.com \
--cc=libc-alpha@sourceware.org \
--cc=patrick.mcgehearty@oracle.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).