[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAH9TF6M2uU1CYe4RL6J=NTYb8mw_wT_OqbS9F8_MDGQnii26_w@mail.gmail.com>
Date: Thu, 29 Aug 2024 15:37:03 +0200
From: Alex Rønne Petersen <alex@...xrp.com>
To: Alexander Monakov <amonakov@...ras.ru>
Cc: musl@...ts.openwall.com
Subject: Re: [PATCH] configure: prevent compilers from turning a * b +
c into fma(a, b, c)
On Wed, Aug 28, 2024 at 11:11 PM Alexander Monakov <amonakov@...ras.ru> wrote:
>
>
> On Wed, 28 Aug 2024, Rich Felker wrote:
>
> > I'm not clear on why the -mno-fused-madd would be needed. It should be
> > a no-op with -ffp-contract=off unless __builtin_fma were called
> > explicitly or something, no?
>
> I think Glibc was applying it as a workaround for old GCC releases where
> target backends were doing unrestricted contraction, or -ffp-contract=
> option did not exist in the first place.
That's right. FWIW: It was added to glibc 12 years ago. I checked
current GCC sources and I can find no indication that the
`-mno-fused-madd` is necessary today; the optimizer looks to be
respecting `-ffp-contract=off`. If we don't care about ancient GCC
versions, I think it'd be completely reasonable to leave out that
flag.
>
> Alexander
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
