From: "Keith Packard" <keithp@keithp.com>
To: Sebastian Huber <sebastian.huber@embedded-brains.de>,
Joseph Myers <joseph@codesourcery.com>
Cc: newlib@sourceware.org
Subject: Re: [PATCH 0/3] ARM with only 32-bit floats do not have fast 64-bit FMA
Date: Tue, 01 Sep 2020 22:35:42 -0700 [thread overview]
Message-ID: <87tuwgaggh.fsf@keithp.com> (raw)
In-Reply-To: <87zh68agy9.fsf@keithp.com>
[-- Attachment #1: Type: text/plain, Size: 1029 bytes --]
"Keith Packard" <keithp@keithp.com> writes:
> That implementation violates the spec though because it does two
> binary operations involving two roundings, so you get a different answer
> than you would with a true fma.
Hrm. C99 and C17 both have macros to detect whether fma is 'fast' or
not: FP_FAST_FMA, FP_FAST_FMAF and FP_FAST_FMAL. This page:
https://en.cppreference.com/w/cpp/numeric/math/fma
has a nice parenthetical comment:
"If ... defined, the function std::fma evaluates faster (in addition to
being more precise) than the expression x*y+z."
If C99 or C17 included 'in addition to being more precise', it would
be much more obvious to me that we should include the fall-back fma
implementation.
So, we should at least change the CPP defines that we have in
math_config.h to match the C99 and C17 specs.
Is it reasonable to assume that applications which care about accuracy
will also be checking these defines and using them as the C++ standard
appears to?
--
-keith
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]
next prev parent reply other threads:[~2020-09-02 5:35 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-08 22:34 Keith Packard
2020-08-08 22:34 ` [PATCH 1/3] libm: ARM without HW double does not have fast FMA Keith Packard
2020-08-08 22:34 ` [PATCH 2/3] libm: Detect fast fmaf support Keith Packard
2020-08-08 22:34 ` [PATCH 3/3] libm/machine/arm: Add optimized fmaf and fma when available Keith Packard
2020-08-10 9:30 ` [PATCH 0/3] ARM with only 32-bit floats do not have fast 64-bit FMA Corinna Vinschen
2020-08-10 14:43 ` Szabolcs Nagy
2020-08-10 15:19 ` Keith Packard
2020-08-10 19:06 ` Corinna Vinschen
2020-09-01 16:32 ` Sebastian Huber
2020-09-01 17:21 ` Sebastian Huber
2020-09-01 18:04 ` Sebastian Huber
2020-09-01 19:28 ` Keith Packard
2020-09-01 21:16 ` Joseph Myers
2020-09-01 23:06 ` Keith Packard
2020-09-02 4:41 ` Sebastian Huber
2020-09-02 5:25 ` Keith Packard
2020-09-02 5:35 ` Keith Packard [this message]
2020-09-02 17:12 ` Joseph Myers
2020-09-02 17:59 ` Sebastian Huber
2020-09-02 20:39 ` Keith Packard
2020-09-01 19:50 ` Keith Packard
2020-09-07 14:09 Eric Bresie
2020-09-07 17:16 ` Keith Packard
2020-09-07 20:16 ` Brian Inglis
2020-09-07 22:23 ` Keith Packard
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=87tuwgaggh.fsf@keithp.com \
--to=keithp@keithp.com \
--cc=joseph@codesourcery.com \
--cc=newlib@sourceware.org \
--cc=sebastian.huber@embedded-brains.de \
/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).