[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <89f8a560-66d7-788-eb8c-96a8f66f119f@esi.com.au>
Date: Sun, 2 Aug 2026 18:58:29 +1000 (AEST)
From: Damian McGuckin <damianm@....com.au>
To: Markus Wichmann <nullplan@....net>
cc: musl@...ts.openwall.com
Subject: Re: Does MUSL assume both float and double always exist?
On Sun, 2 Aug 2026, Markus Wichmann wrote:
>> IEEE 754 only mandates one. And it can be neither of float, double or long
>> double. Again irrelevant. These days it can be BINARY16.
>
> ISO-C Annex F requires that C implementations conforming to IEEE 754
> have float be binary32 and double be binary64 format. So no, it cannot
> be binary16. GCC implements binary16 with the __fp16 extension type that
> musl does not support.
My comment was purely with regard to IEEE 754 in isolation although the
subject was all about having float and double types (where the latter had
to be more accurate than the former).
A C compiler which implements only binary16 is IEEE 754 (ISO 60559)
compliant or will be with the 2029 release. However, it is not compliant
with ISO 9899 (ISO-C). Annex F does not mention an ISO-C _float16 type
currently (although Annex H does).
Thanks - Damian
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.
