From: Uros Bizjak <ubizjak@gmail.com>
To: Ilya Tocar <tocarip.intel@gmail.com>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: [PATCH, i386, testsuite] FMA intrinsics
Date: Mon, 22 Aug 2011 20:41:00 -0000 [thread overview]
Message-ID: <CAFULd4ZZ6ozxao7OqXh=GetQr6okBSJXiYUB+io4egrc-KuChA@mail.gmail.com> (raw)
In-Reply-To: <CAFFGCRB-auRYMdLGPzSdb9ftGd++vtgeWzOG2xARQwe4vij9iQ@mail.gmail.com>
On Mon, Aug 22, 2011 at 6:25 PM, Ilya Tocar <tocarip.intel@gmail.com> wrote:
>> You don't need to add "negated" versions, one FMA builtin per mode is
>> enough, please see existing FMA4 descriptions. Just put unary minus
>> sign in the intrinsics header for "negated" operand and let GCC do its
>> job. Please see existing FMA4 intrinsics header.
>>
> Actually i tried that.But in such case when i compile(FMA4 example)
> #include <x86intrin.h>
> extern __m128 a,b,c;
> void foo(){
> a = _mm_nmsub_ps(a,b,c);
> }
> with -S -O0 -mfma4
> The asm have
>
> vxorps %xmm1, %xmm0, %xmm0
> vmovaps -16(%rbp), %xmm1
> vmovaps .LC0(%rip), %xmm2
> vxorps %xmm2, %xmm1, %xmm1
> vfmaddps %xmm0, -32(%rbp), %xmm1, %xmm0
> So vfmaddps of negated values is generated instead of vfnmsubps.
> I think it is bad that intrinsic for instruction can generate code
> without this instruction.
> So to make sure that exact instruction is always generated i
> introduced additional expands and builtins.
> Is it wrong?
This is artificial limitation. User requested the functionality of the
intrinsic, and should not bother with how the compiler realizes it.
With -O2, negation would propagate into the insn during combine pass,
and optimal instruction would be generated.
So, to answer your question - it is wrong to expect exact instruction
from builtins. Maybe from using -O0, but this should not be used
anyway in the testsuite.
Uros.
next prev parent reply other threads:[~2011-08-22 19:26 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-08-20 11:23 Uros Bizjak
2011-08-22 17:32 ` Ilya Tocar
2011-08-22 20:41 ` Uros Bizjak [this message]
2011-08-23 14:55 ` Ilya Tocar
2011-08-23 15:33 ` Uros Bizjak
2011-08-24 10:06 ` Ilya Tocar
2011-08-24 10:12 ` Jakub Jelinek
2011-08-24 11:12 ` Ilya Tocar
2011-08-24 13:31 ` Uros Bizjak
2011-08-24 13:52 ` Ilya Tocar
2011-08-24 22:00 ` Uros Bizjak
2011-08-25 10:15 ` Ilya Tocar
2011-08-25 10:46 ` Uros Bizjak
2011-08-25 11:46 ` Ilya Tocar
2011-08-25 11:46 ` Jakub Jelinek
2011-08-25 11:49 ` Ilya Tocar
2011-08-26 11:03 ` Ilya Tocar
2011-08-26 14:07 ` H.J. Lu
2011-08-26 15:47 ` Ilya Tocar
2011-08-26 17:02 ` H.J. Lu
2011-08-26 17:06 ` H.J. Lu
2011-08-30 12:05 ` Ilya Tocar
2011-08-30 14:02 ` H.J. Lu
2011-08-30 14:10 ` Ilya Tocar
2011-08-30 14:32 ` Ilya Tocar
2011-08-30 15:25 ` H.J. Lu
-- strict thread matches above, loose matches on Subject: below --
2011-08-17 14:08 [PATCH, i386, testsuite]FMA intrinsics Ilya Tocar
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='CAFULd4ZZ6ozxao7OqXh=GetQr6okBSJXiYUB+io4egrc-KuChA@mail.gmail.com' \
--to=ubizjak@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=tocarip.intel@gmail.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).