<html>
<head>
<base href="https://bugs.llvm.org/">
</head>
<body><table border="1" cellspacing="0" cellpadding="8">
<tr>
<th>Bug ID</th>
<td><a class="bz_bug_link
bz_status_NEW "
title="NEW - Using both -x cuda and -fopenmp complex wrappers block compilation"
href="https://bugs.llvm.org/show_bug.cgi?id=48014">48014</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>Using both -x cuda and -fopenmp complex wrappers block compilation
</td>
</tr>
<tr>
<th>Product</th>
<td>clang
</td>
</tr>
<tr>
<th>Version</th>
<td>11.0
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>All
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>normal
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>OpenMP
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedclangbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>joachim@joameyer.de
</td>
</tr>
<tr>
<th>CC</th>
<td>llvm-bugs@lists.llvm.org
</td>
</tr></table>
<p>
<div>
<pre>This is very related to <a class="bz_bug_link
bz_status_RESOLVED bz_closed"
title="RESOLVED FIXED - CUDA and OpenMP math functions conflicts (using both -xcuda and -fopenmp)"
href="show_bug.cgi?id=45533">Bug 45533</a>.
If compiling with both -x cuda and -fopenmp (the latter to target CPU), a
number of errors similar to the following are emitted:
In file included from <built-in>:1:
In file included from
/home/joachim/Projekte/sycl/install/llvm11/lib/clang/11.0.0/include/__clang_cuda_runtime_wrapper.h:406:
/home/joachim/Projekte/sycl/install/llvm11/lib/clang/11.0.0/include/__clang_cuda_complex_builtins.h:76:7:
error: no matching function for call to '__nv_isnand'
if (_ISNANd(__real__(z)) && _ISNANd(__imag__(z))) {
^~~~~~~
/home/joachim/Projekte/sycl/install/llvm11/lib/clang/11.0.0/include/__clang_cuda_complex_builtins.h:47:17:
note: expanded from macro '_ISNANd'
#define _ISNANd __nv_isnand
^~~~~~~~~~~
/home/joachim/Projekte/sycl/install/llvm11/lib/clang/11.0.0/include/__clang_cuda_libdevice_declares.h:226:16:
note: candidate function not viable: call to __device__ function from __host__
function
__DEVICE__ int __nv_isnand(double __a);
This is fixed in <a href="https://reviews.llvm.org/D90415">https://reviews.llvm.org/D90415</a> for master.
As this is the only missing piece to get the combination of -x cuda and
-fopenmp running with clang 11, I would suggest to backport the change to
release/11.x as well.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are on the CC list for the bug.</li>
</ul>
</body>
</html>