diff options
| author | Adhemerval Zanella <adhemerval.zanella@linaro.org> | 2019-01-11 17:31:17 +0000 |
|---|---|---|
| committer | Adhemerval Zanella <adhemerval.zanella@linaro.org> | 2019-01-11 17:31:17 +0000 |
| commit | 7aafc4d66a52bf0c6f03db4b805b844b3c79dffa (patch) | |
| tree | d2a229ae28b7ed56abc30aa46d3202f42c1b0740 | |
| parent | ef48e7bccee4f956744564e47b9c932e0605f7a8 (diff) | |
[libcxx] Call __count_bool_true for bitset count
This patch aims to help clang with better information so it can inline
__bit_reference count function usage for both std::biset. Current clang
inliner can not infer that the passed typed will be used only to select
the optimized variant, it evaluates the type argument and type check as
a load plus compare (although later optimization phases correctly
optimized this out).
It is mainly to help llvm inliner to generate better code for std::bitset
count for aarch64. It helps on both runtime and code size, since if inline
decides that _VSTD::count should not be inlined the vectorization will
create both aligned and unaligned variants (which add both code size and
runtime costs)
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@350936 91177308-0d34-0410-b5e6-96231b3b80d8
| -rw-r--r-- | include/bitset | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/include/bitset b/include/bitset index 6e28596d0..98947e027 100644 --- a/include/bitset +++ b/include/bitset @@ -991,7 +991,7 @@ inline size_t bitset<_Size>::count() const _NOEXCEPT { - return static_cast<size_t>(_VSTD::count(base::__make_iter(0), base::__make_iter(_Size), true)); + return static_cast<size_t>(__count_bool_true(base::__make_iter(0), _Size)); } template <size_t _Size> |