From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <sourceware-bugzilla@sourceware.org>
Received: by sourceware.org (Postfix, from userid 48)
	id 7DC733858CD1; Mon, 18 Mar 2024 22:17:56 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 7DC733858CD1
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org;
	s=default; t=1710800276;
	bh=ggqhFoIJRjNxWxVtWo5lo00u0M3qDPfDoiBZ6znmybE=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=BONTucqfTHObLatuWeaZrYd5hPODQdOX9MePAUreBtWdPs09HVmxBCPAyKc1h50G0
	 pnIKaFsyj3fd2kZ7VaNgvoVWxEWOAxPQl6OjRhnpHgl2YULeX9ziHb+17oYNSGYSZF
	 mg8rGJOI+KYv/k4fqntQrJP6Y+YfI2HwVnXZsggc=
From: "wdijkstr at arm dot com" <sourceware-bugzilla@sourceware.org>
To: glibc-bugs@sourceware.org
Subject: [Bug math/28472] pow(10, i) accuracy
Date: Mon, 18 Mar 2024 22:17:56 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: glibc
X-Bugzilla-Component: math
X-Bugzilla-Version: 2.31
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: wdijkstr at arm dot com
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at sourceware dot org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-28472-131-HLgFT9HvUY@http.sourceware.org/bugzilla/>
In-Reply-To: <bug-28472-131@http.sourceware.org/bugzilla/>
References: <bug-28472-131@http.sourceware.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://sourceware.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <glibc-bugs.sourceware.org>

https://sourceware.org/bugzilla/show_bug.cgi?id=3D28472

--- Comment #22 from Wilco <wdijkstr at arm dot com> ---
(In reply to b. from comment #21)
> dear guys,=20=20
>=20=20=20
> in a way you are talking this point to death,=20=20
>=20=20=20
> this issue is not about powers of two, usually they have=20=20=20
> few problems in binary datatypes, but about powers of 10!=20=20
>=20=20=20
> it's important to have them exact ( as good as possible )=20=20
> at integral powers of ten,=20=20

We do get all of the powers of 10 that are exactly representable correct. A=
nd
almost all of the inexact cases are rounded correctly.

> and regarding monotonity it's surely better to smoothe the=20=20
> nearby values to the exact values than the other way.=20=20
>=20=20=20
> Note, we need 2! results changed for binary64's ( bin80's=20=20
> are worse ) and there are no monotonity issues around them.

There isn't an obvious fix for the current implementation. If you can find a
way without a performance hit, I'd love to hear it. If you want correctly
rounded results (and don't care about performance), just use a correctly
rounded math library.

--=20
You are receiving this mail because:
You are on the CC list for the bug.=