From: Segher Boessenkool <segher@kernel.crashing.org>
To: Richard Biener <richard.guenther@gmail.com>
Cc: HAO CHEN GUI <guihaoc@linux.ibm.com>,
gcc-patches <gcc-patches@gcc.gnu.org>,
Bill Schmidt <wschmidt@linux.ibm.com>
Subject: Re: [PATCH, rs6000] Disable gimple fold for float or double vec_minmax when fast-math is not set
Date: Wed, 13 Oct 2021 13:19:27 -0500 [thread overview]
Message-ID: <20211013181927.GA614@gate.crashing.org> (raw)
In-Reply-To: <CAFiYyc0YdUbba8eVPkL76G8boUP+Td+OoiVt45BA4imfFa0-gg@mail.gmail.com>
On Wed, Oct 13, 2021 at 10:29:26AM +0200, Richard Biener wrote:
> On Wed, Oct 13, 2021 at 9:43 AM HAO CHEN GUI <guihaoc@linux.ibm.com> wrote:
> > As to IEEE behavior, do you mean "Minimum and maximum operations" defined in IEEE-754 2019? If so, I think VSX/altivec min/max instructions don't conform with it. It demands a quite NaN if either operand is a NaN while our instructions don't.
> >
> > IEEE-754 2019 maximum(x, y) is xif x>y, yif y>x, and a quiet NaN if either operand is a NaN, according to 6.2. For this operation, +0 compares greater than −0. Otherwise (i.e., when x=y and signs are the same) it is either xor y. Actions for xvmaxdp
>
> Hmm, then I do not understand the reason for the patch - people using
> the intrinsics cannot expect IEEE semantics then.
> So you are concerned that people don't get the 1:1 machine instruction
> but eventually the IEEE conforming MIN/MAX_EXPR?
I do not know about Gimple MIN_EXPR (it is not documented?), but the
RTL "smin" is meaningless in the presence of NaNs or signed zeros. This
is documented (in rtl.texi):
"""
@findex smin
@findex smax
@cindex signed minimum
@cindex signed maximum
@item (smin:@var{m} @var{x} @var{y})
@itemx (smax:@var{m} @var{x} @var{y})
Represents the smaller (for @code{smin}) or larger (for @code{smax}) of
@var{x} and @var{y}, interpreted as signed values in mode @var{m}.
When used with floating point, if both operands are zeros, or if either
operand is @code{NaN}, then it is unspecified which of the two operands
is returned as the result.
"""
(not exactly meaningless, okay, but not usable for almost anything).
> But that can then still happen with -ffast-math so I wonder what's the point.
That :-)
Segher
next prev parent reply other threads:[~2021-10-13 18:20 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-12 8:57 HAO CHEN GUI
2021-10-12 9:57 ` Richard Biener
2021-10-13 7:43 ` HAO CHEN GUI
2021-10-13 8:29 ` Richard Biener
2021-10-13 9:15 ` HAO CHEN GUI
2021-10-13 18:19 ` Segher Boessenkool [this message]
2021-10-13 22:04 ` Joseph Myers
2021-10-13 18:28 ` Segher Boessenkool
-- strict thread matches above, loose matches on Subject: below --
2021-11-01 6:48 HAO CHEN GUI
2021-11-02 0:21 ` David Edelsohn
2021-11-02 2:40 ` HAO CHEN GUI
2021-11-02 13:12 ` David Edelsohn
2021-11-03 2:14 ` HAO CHEN GUI
2021-10-20 9:04 HAO CHEN GUI
2021-10-20 16:19 ` Segher Boessenkool
2021-10-21 6:25 ` HAO CHEN GUI
2021-10-13 15:02 David Edelsohn
2021-08-24 8:52 HAO CHEN GUI
2021-08-24 20:04 ` Bill Schmidt
2021-08-24 22:40 ` Segher Boessenkool
2021-08-25 7:06 ` HAO CHEN GUI
2021-08-25 7:50 ` Kewen.Lin
2021-08-25 8:17 ` HAO CHEN GUI
2021-08-25 8:44 ` HAO CHEN GUI
2021-08-25 12:34 ` Bill Schmidt
2021-08-26 1:19 ` HAO CHEN GUI
2021-09-06 6:01 ` HAO CHEN GUI
2021-10-11 15:55 ` Segher Boessenkool
2021-10-11 16:07 ` Segher Boessenkool
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=20211013181927.GA614@gate.crashing.org \
--to=segher@kernel.crashing.org \
--cc=gcc-patches@gcc.gnu.org \
--cc=guihaoc@linux.ibm.com \
--cc=richard.guenther@gmail.com \
--cc=wschmidt@linux.ibm.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).