Skip to content

vector_algorithms.cpp: minmax for 64-bit elements: replace ugly x86 workaround with a nice one #4661

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
May 20, 2024
Merged

vector_algorithms.cpp: minmax for 64-bit elements: replace ugly x86 workaround with a nice one #4661

Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
455a184
Do stack spill; compiler optimizes it away
AlexGuteniev May 7, 2024
00dde4c
Another occurrence!
AlexGuteniev May 7, 2024
b702589
consume less stack (virtually)
AlexGuteniev May 7, 2024
a2798f6
Comment grammar.
StephanTLavavej May 12, 2024
5d85e7e
`_Minmax_traits_8`: Implement `_Get_any` with `_Get_v_pos` and `stati…
StephanTLavavej May 12, 2024
39babc0
`_Minmax_traits_d::_Get_v_pos` is identical to `_Minmax_traits_8::_Ge…
StephanTLavavej May 12, 2024
5563c6e
`_Minmax_traits_d::_Get_any_u` can punch through to `_Minmax_traits_8…
StephanTLavavej May 12, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 5 additions & 16 deletions stl/src/vector_algorithms.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -901,13 +901,9 @@ namespace {
}

static _Signed_t _Get_any(const __m128i _Cur) noexcept {
#ifdef _M_IX86
return static_cast<_Signed_t>(
(static_cast<_Unsigned_t>(static_cast<uint32_t>(_mm_extract_epi32(_Cur, 1))) << 32)
| static_cast<_Unsigned_t>(static_cast<uint32_t>(_mm_cvtsi128_si32(_Cur))));
#else // ^^^ x86 / x64 vvv
return static_cast<_Signed_t>(_mm_cvtsi128_si64(_Cur));
#endif // ^^^ x64 ^^^
// With optimizations enabled, compiles into register movement, rather than an actual stack spill.
// Works around the absence of _mm_cvtsi128_si64 on 32-bit.
return static_cast<_Signed_t>(_Get_v_pos(_Cur, 0));
}

static _Unsigned_t _Get_v_pos(const __m128i _Idx, const unsigned long _H_pos) noexcept {
Expand Down Expand Up @@ -1114,18 +1110,11 @@ namespace {
}

static uint64_t _Get_any_u(const __m128i _Cur) noexcept {
#ifdef _M_IX86
return (static_cast<uint64_t>(static_cast<uint32_t>(_mm_extract_epi32(_Cur, 1))) << 32)
| static_cast<uint64_t>(static_cast<uint32_t>(_mm_cvtsi128_si32(_Cur)));
#else // ^^^ x86 / x64 vvv
return static_cast<uint64_t>(_mm_cvtsi128_si64(_Cur));
#endif // ^^^ x64 ^^^
return _Minmax_traits_8::_Get_v_pos(_Cur, 0);
}

static uint64_t _Get_v_pos(const __m128i _Idx, const unsigned long _H_pos) noexcept {
uint64_t _Array[2];
_mm_storeu_si128(reinterpret_cast<__m128i*>(&_Array), _Idx);
return _Array[_H_pos >> 3];
return _Minmax_traits_8::_Get_v_pos(_Idx, _H_pos);
}

static __m128d _Cmp_eq(const __m128d _First, const __m128d _Second) noexcept {
Expand Down