From: "Dominique d'Humières" <dominiq@lps.ens.fr>
To: sgk@troutmask.apl.washington.edu
Cc: Janus Weil <janus@gcc.gnu.org>, gfortran <fortran@gcc.gnu.org>
Subject: Re: [Patch, Fortran] PR 82143: add a -fdefault-real-16 flag
Date: Fri, 22 Sep 2017 22:46:00 -0000 [thread overview]
Message-ID: <C544F65C-D534-4908-9F76-64DBDE81EF75@lps.ens.fr> (raw)
In-Reply-To: <20170922143709.GB17729@troutmask.apl.washington.edu>
> test_fpu is a horrible benchmark for testing floating point performance.
Why? It is mostly testing a tiny subset of LAPACK subroutines and yes know that linear algebra does not cover all the uses of floating point calculations. What I have done (almost six years ago) in the REAL(16) variant was to replace all the DOUBLE PRECISION with REAL(16) and a few DXXX with XXX (as in DABS to ABS), the later preventing me to use -fdefault-real-8.
Note that I gave some timing for other tests in the polyhedron suite. If someone is interested I can provide the following variants for REAL(16): capacita_16.f90, channel_16.f90, fatigue_16.f90, nduct_16.f90, kepler_16.f90, linpk_16.f90, nf_16.f90, protein_16.f90, test_fpu_16.f90, and tfft_16.f90.
> How big is your L1 and L2 cache?
Although I don’t think it is relevant, I am using a laptop with a 2.8 Ghz Core i7, 4 cores/8 threads. AFAIK
L1 is 64 kbytes per core (likely 32k instructions, 32k data),
L2 is 256 kbytes per core,
L3 is at least 6 Mbytes (may be 8, don’t remember).
> That aside, no one who does numerical computations for a living
> will ever confuse a software-backed implementation of REAl(16)
> as being as fast as a hardware-backed implementation of REAL.
What let you think I am doing the confusion?
> If a calculation requires 113-bits of precision, it does not
> matter how fast a 24-bit floating point type is.
But how slow quadmath is really matter.
>
> --
> Steve
> 20170425 https://www.youtube.com/watch?v=VWUpyCsUKR4
> 20161221 https://www.youtube.com/watch?v=IbCHE-hONow
Dominique
next prev parent reply other threads:[~2017-09-22 22:46 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-09-18 9:31 Dominique d'Humières
2017-09-18 18:13 ` Janus Weil
2017-09-22 10:06 ` Dominique d'Humières
2017-09-22 11:21 ` Janus Weil
2017-09-22 14:37 ` Steve Kargl
2017-09-22 22:46 ` Dominique d'Humières [this message]
2017-09-22 23:47 ` Steve Kargl
2017-12-03 17:36 ` Gerald Pfeifer
-- strict thread matches above, loose matches on Subject: below --
2017-09-25 15:07 David Edelsohn
2017-09-25 21:14 ` Janus Weil
2017-09-25 21:23 ` Steve Kargl
2017-09-26 8:44 ` Janus Weil
2017-09-26 9:03 ` Janus Weil
2017-09-26 9:12 ` Rainer Orth
2017-09-26 9:47 ` Janus Weil
2017-09-26 14:54 ` David Edelsohn
2017-09-18 9:38 Dominique d'Humières
2017-09-17 20:42 Janus Weil
2017-09-18 4:51 ` Steve Kargl
2017-09-18 7:02 ` Janus Weil
2017-09-18 14:08 ` Steve Kargl
2017-09-18 17:57 ` Janus Weil
2017-09-21 7:10 ` Janus Weil
2017-09-21 20:38 ` Steve Kargl
2017-09-22 5:03 ` Janus Weil
2017-09-22 7:12 ` Janne Blomqvist
2017-09-22 9:44 ` Janus Weil
2017-09-22 19:32 ` Janus Weil
2017-09-23 13:19 ` Janus Weil
2017-09-22 14:06 ` Steve Kargl
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=C544F65C-D534-4908-9F76-64DBDE81EF75@lps.ens.fr \
--to=dominiq@lps.ens.fr \
--cc=fortran@gcc.gnu.org \
--cc=janus@gcc.gnu.org \
--cc=sgk@troutmask.apl.washington.edu \
/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).