1berry/qtbase
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix for ClangCl missing F16C intrinsics in qfloat16.h

Browse Source 
Done-with: Thiago Macieira <thiago.macieira@intel.com>
Fixes: QTBUG-101175
Change-Id: I46ef5c473f174e54a8cffffd16e5334d3b5944b8
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
This commit is contained in:
Mario Emmenlauer 2022-03-07 14:48:32 +01:00 committed by Thiago Macieira
parent 860b8d00ca
commit f3640d9086
1 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/corelib/global/qfloat16.h
View File

@ -48,12 +48,13 @@
#include <string.h>
#if defined(QT_COMPILER_SUPPORTS_F16C) && defined(__AVX2__) && !defined(__F16C__)
// All processors that support AVX2 do support F16C too. That doesn't mean
// we're allowed to use the intrinsics directly, so we'll do it only for
// the Intel and Microsoft's compilers.
# if defined(Q_CC_INTEL) || defined(Q_CC_MSVC)
// All processors that support AVX2 do support F16C too, so we could enable the
// feature unconditionally if __AVX2__ is defined. However, all currently
// supported compilers except Microsoft's are able to define __F16C__ on their
// own when the user enables the feature, so we'll trust them.
# if defined(Q_CC_MSVC) && !defined(Q_CC_CLANG)
# define __F16C__ 1
# endif
# endif
#endif
#if defined(QT_COMPILER_SUPPORTS_F16C) && defined(__F16C__)