Skip to content
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

fix: Disable optimization to avoid pontential errors #640

Merged
merged 1 commit into from
Jun 26, 2024
Merged

fix: Disable optimization to avoid pontential errors #640

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
26 changes: 19 additions & 7 deletions sse2neon.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -106,6 +106,17 @@
#pragma message("Macro name collisions may happen with unsupported compilers.")
#endif


#if defined(__GNUC__) && !defined(__clang__)
#pragma push_macro("FORCE_INLINE_OPTNONE")
#define FORCE_INLINE_OPTNONE static inline __attribute__((optimize("O0")))
#elif defined(__clang__)
#pragma push_macro("FORCE_INLINE_OPTNONE")
#define FORCE_INLINE_OPTNONE static inline __attribute__((optnone))
#else
#define FORCE_INLINE_OPTNONE FORCE_INLINE
#endif

#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ < 10
#warning "GCC versions earlier than 10 are not supported."
#endif
Expand Down Expand Up @@ -579,8 +590,8 @@ FORCE_INLINE __m128d _mm_ceil_pd(__m128d);
FORCE_INLINE __m128 _mm_ceil_ps(__m128);
FORCE_INLINE __m128d _mm_floor_pd(__m128d);
FORCE_INLINE __m128 _mm_floor_ps(__m128);
FORCE_INLINE __m128d _mm_round_pd(__m128d, int);
FORCE_INLINE __m128 _mm_round_ps(__m128, int);
FORCE_INLINE_OPTNONE __m128d _mm_round_pd(__m128d, int);
FORCE_INLINE_OPTNONE __m128 _mm_round_ps(__m128, int);
// SSE4.2
FORCE_INLINE uint32_t _mm_crc32_u8(uint32_t, uint8_t);

Expand Down Expand Up @@ -2162,7 +2173,7 @@ FORCE_INLINE int _mm_movemask_ps(__m128 a)
// Multiply packed single-precision (32-bit) floating-point elements in a and b,
// and store the results in dst.
// https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_mul_ps
FORCE_INLINE __m128 _mm_mul_ps(__m128 a, __m128 b)
FORCE_INLINE_OPTNONE __m128 _mm_mul_ps(__m128 a, __m128 b)
{
return vreinterpretq_m128_f32(
vmulq_f32(vreinterpretq_f32_m128(a), vreinterpretq_f32_m128(b)));
Expand Down Expand Up @@ -3843,7 +3854,7 @@ FORCE_INLINE __m128 _mm_cvtepi32_ps(__m128i a)
// Convert packed double-precision (64-bit) floating-point elements in a to
// packed 32-bit integers, and store the results in dst.
// https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtpd_epi32
FORCE_INLINE __m128i _mm_cvtpd_epi32(__m128d a)
FORCE_INLINE_OPTNONE __m128i _mm_cvtpd_epi32(__m128d a)
{
// vrnd32xq_f64 not supported on clang
#if defined(__ARM_FEATURE_FRINT) && !defined(__clang__)
Expand All @@ -3862,7 +3873,7 @@ FORCE_INLINE __m128i _mm_cvtpd_epi32(__m128d a)
// Convert packed double-precision (64-bit) floating-point elements in a to
// packed 32-bit integers, and store the results in dst.
// https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_cvtpd_pi32
FORCE_INLINE __m64 _mm_cvtpd_pi32(__m128d a)
FORCE_INLINE_OPTNONE __m64 _mm_cvtpd_pi32(__m128d a)
{
__m128d rnd = _mm_round_pd(a, _MM_FROUND_CUR_DIRECTION);
double d0 = ((double *) &rnd)[0];
Expand Down Expand Up @@ -7421,7 +7432,7 @@ FORCE_INLINE __m128i _mm_packus_epi32(__m128i a, __m128i b)
// the rounding parameter, and store the results as packed double-precision
// floating-point elements in dst.
// https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm_round_pd
FORCE_INLINE __m128d _mm_round_pd(__m128d a, int rounding)
FORCE_INLINE_OPTNONE __m128d _mm_round_pd(__m128d a, int rounding)
{
#if defined(__aarch64__) || defined(_M_ARM64)
switch (rounding) {
Expand Down Expand Up @@ -7490,7 +7501,7 @@ FORCE_INLINE __m128d _mm_round_pd(__m128d a, int rounding)
// the rounding parameter, and store the results as packed single-precision
// floating-point elements in dst.
// software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_mm_round_ps
FORCE_INLINE __m128 _mm_round_ps(__m128 a, int rounding)
FORCE_INLINE_OPTNONE __m128 _mm_round_ps(__m128 a, int rounding)
{
#if (defined(__aarch64__) || defined(_M_ARM64)) || \
defined(__ARM_FEATURE_DIRECTED_ROUNDING)
Expand Down Expand Up @@ -9280,6 +9291,7 @@ FORCE_INLINE uint64_t _rdtsc(void)
#if defined(__GNUC__) || defined(__clang__)
#pragma pop_macro("ALIGN_STRUCT")
#pragma pop_macro("FORCE_INLINE")
#pragma pop_macro("FORCE_INLINE_OPTNONE")
#endif

#if defined(__GNUC__) && !defined(__clang__)
Expand Down
Loading