Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 29 Oct 2019 18:05:48 +0000
From:      bugzilla-noreply@freebsd.org
To:        toolchain@FreeBSD.org
Subject:   [Bug 241550] Base Clang can't compile trivial CUDA programs: error: no matching function for call to '__isnan'
Message-ID:  <bug-241550-29464-ogQCVICRb6@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-241550-29464@https.bugs.freebsd.org/bugzilla/>
References:  <bug-241550-29464@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D241550

Dimitry Andric <dim@FreeBSD.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |theraven@FreeBSD.org

--- Comment #2 from Dimitry Andric <dim@FreeBSD.org> ---
It's unfortunate that the cuda headers use these __isnan functions, which s=
eem
to be glibc specific.

That said, I see the error is not in the clang-supplied cuda wrapper header=
s,
but in
/usr/home/arr/cuda/var/cuda-repo-10-0-local-10.0.130-410.48/usr/local/cuda-=
10.0//include/crt/math_functions.hpp.
 So this is some specific thing that is (maybe) being used all of the place=
 in
CUDA code?

We do indeed have an __inline_isnan, but it is currently only used for the
__fp_type_select() macro in math.h.  The name originates in base r253215 by
David, which says "Cleanups to math.h that prevent namespace conflicts with
C++".

It looks like the name used to be __isnan before that commit, but why it was
changed is unclear.  There is no __isnan in the C++ standards, as far as I =
can
see.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-241550-29464-ogQCVICRb6>