This is an archive of the discontinued LLVM Phabricator instance.

Differential D21988

[AVX512] minor fix in sqrt{ss|sd} intrinsics arguments
ClosedPublic

Authored by AsafBadouh on Jul 5 2016, 4:30 AM.

Details

Reviewers
guyblank
delena
m_zuckerman
Commits
rGf9cdb8de7ab3: [AVX512] minor fix in sqrt{ss|sd} intrinsics arguments
rC274541: [AVX512] minor fix in sqrt{ss|sd} intrinsics arguments
rL274541: [AVX512] minor fix in sqrt{ss|sd} intrinsics arguments

Diff Detail

Repository
rL LLVM

Event Timeline

AsafBadouh updated this revision to Diff 62734.Jul 5 2016, 4:30 AM
AsafBadouh retitled this revision from to [AVX512] minor fix in sqrt{ss|sd} intrinsics arguments.
AsafBadouh updated this object.
AsafBadouh added reviewers: m_zuckerman, guyblank, delena.
AsafBadouh set the repository for this revision to rL LLVM.
AsafBadouh added a subscriber: cfe-commits.
guyblank accepted this revision.Jul 5 2016, 4:36 AM
guyblank edited edge metadata.
This revision is now accepted and ready to land.Jul 5 2016, 4:36 AM
Closed by commit rL274541: [AVX512] minor fix in sqrt{ss|sd} intrinsics arguments (authored by abadouh). · Explain WhyJul 5 2016, 4:43 AM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
cfe/
trunk/
lib/
Headers/
36 lines

Diff 62735

cfe/trunk/lib/Headers/avx512fintrin.h

  • This file is larger than 256 KB, so syntax highlighting is disabled by default.
Show First 20 Lines • Show All 7,202 Lines • ▼ Show 20 Lines
#define _mm512_maskz_shuffle_ps(U, M, V, imm) __extension__ ({ \ #define _mm512_maskz_shuffle_ps(U, M, V, imm) __extension__ ({ \
(__m512)__builtin_ia32_shufps512_mask((__v16sf)(__m512)(M), \ (__m512)__builtin_ia32_shufps512_mask((__v16sf)(__m512)(M), \
(__v16sf)(__m512)(V), (int)(imm), \ (__v16sf)(__m512)(V), (int)(imm), \
(__v16sf)_mm512_setzero_ps(), \ (__v16sf)_mm512_setzero_ps(), \
(__mmask16)(U)); }) (__mmask16)(U)); })
#define _mm_sqrt_round_sd(A, B, R) __extension__ ({ \ #define _mm_sqrt_round_sd(A, B, R) __extension__ ({ \
(__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(B), \ (__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(A), \ (__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \ (__v2df)_mm_setzero_pd(), \
(__mmask8)-1, (int)(R)); }) (__mmask8)-1, (int)(R)); })
static __inline__ __m128d __DEFAULT_FN_ATTRS static __inline__ __m128d __DEFAULT_FN_ATTRS
_mm_mask_sqrt_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) _mm_mask_sqrt_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
{ {
return (__m128d) __builtin_ia32_sqrtsd_round_mask ( (__v2df) __B, return (__m128d) __builtin_ia32_sqrtsd_round_mask ( (__v2df) __A,
(__v2df) __A, (__v2df) __B,
(__v2df) __W, (__v2df) __W,
(__mmask8) __U, (__mmask8) __U,
_MM_FROUND_CUR_DIRECTION); _MM_FROUND_CUR_DIRECTION);
} }
#define _mm_mask_sqrt_round_sd(W, U, A, B, R) __extension__ ({ \ #define _mm_mask_sqrt_round_sd(W, U, A, B, R) __extension__ ({ \
(__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(B), \ (__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(A), \ (__v2df)(__m128d)(B), \
(__v2df)(__m128d)(W), \ (__v2df)(__m128d)(W), \
(__mmask8)(U), (int)(R)); }) (__mmask8)(U), (int)(R)); })
static __inline__ __m128d __DEFAULT_FN_ATTRS static __inline__ __m128d __DEFAULT_FN_ATTRS
_mm_maskz_sqrt_sd (__mmask8 __U, __m128d __A, __m128d __B) _mm_maskz_sqrt_sd (__mmask8 __U, __m128d __A, __m128d __B)
{ {
return (__m128d) __builtin_ia32_sqrtsd_round_mask ( (__v2df) __B, return (__m128d) __builtin_ia32_sqrtsd_round_mask ( (__v2df) __A,
(__v2df) __A, (__v2df) __B,
(__v2df) _mm_setzero_pd (), (__v2df) _mm_setzero_pd (),
(__mmask8) __U, (__mmask8) __U,
_MM_FROUND_CUR_DIRECTION); _MM_FROUND_CUR_DIRECTION);
} }
#define _mm_maskz_sqrt_round_sd(U, A, B, R) __extension__ ({ \ #define _mm_maskz_sqrt_round_sd(U, A, B, R) __extension__ ({ \
(__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(B), \ (__m128d)__builtin_ia32_sqrtsd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(A), \ (__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \ (__v2df)_mm_setzero_pd(), \
(__mmask8)(U), (int)(R)); }) (__mmask8)(U), (int)(R)); })
#define _mm_sqrt_round_ss(A, B, R) __extension__ ({ \ #define _mm_sqrt_round_ss(A, B, R) __extension__ ({ \
(__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(B), \ (__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(A), \ (__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \ (__v4sf)_mm_setzero_ps(), \
(__mmask8)-1, (int)(R)); }) (__mmask8)-1, (int)(R)); })
static __inline__ __m128 __DEFAULT_FN_ATTRS static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_mask_sqrt_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) _mm_mask_sqrt_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
{ {
return (__m128) __builtin_ia32_sqrtss_round_mask ( (__v4sf) __B, return (__m128) __builtin_ia32_sqrtss_round_mask ( (__v4sf) __A,
(__v4sf) __A, (__v4sf) __B,
(__v4sf) __W, (__v4sf) __W,
(__mmask8) __U, (__mmask8) __U,
_MM_FROUND_CUR_DIRECTION); _MM_FROUND_CUR_DIRECTION);
} }
#define _mm_mask_sqrt_round_ss(W, U, A, B, R) __extension__ ({ \ #define _mm_mask_sqrt_round_ss(W, U, A, B, R) __extension__ ({ \
(__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(B), \ (__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(A), \ (__v4sf)(__m128)(B), \
(__v4sf)(__m128)(W), (__mmask8)(U), \ (__v4sf)(__m128)(W), (__mmask8)(U), \
(int)(R)); }) (int)(R)); })
static __inline__ __m128 __DEFAULT_FN_ATTRS static __inline__ __m128 __DEFAULT_FN_ATTRS
_mm_maskz_sqrt_ss (__mmask8 __U, __m128 __A, __m128 __B) _mm_maskz_sqrt_ss (__mmask8 __U, __m128 __A, __m128 __B)
{ {
return (__m128) __builtin_ia32_sqrtss_round_mask ( (__v4sf) __A, return (__m128) __builtin_ia32_sqrtss_round_mask ( (__v4sf) __A,
(__v4sf) __B, (__v4sf) __B,
(__v4sf) _mm_setzero_ps (), (__v4sf) _mm_setzero_ps (),
(__mmask8) __U, (__mmask8) __U,
_MM_FROUND_CUR_DIRECTION); _MM_FROUND_CUR_DIRECTION);
} }
#define _mm_maskz_sqrt_round_ss(U, A, B, R) __extension__ ({ \ #define _mm_maskz_sqrt_round_ss(U, A, B, R) __extension__ ({ \
(__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(B), \ (__m128)__builtin_ia32_sqrtss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(A), \ (__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \ (__v4sf)_mm_setzero_ps(), \
(__mmask8)(U), (int)(R)); }) (__mmask8)(U), (int)(R)); })
static __inline__ __m512 __DEFAULT_FN_ATTRS static __inline__ __m512 __DEFAULT_FN_ATTRS
_mm512_broadcast_f32x4 (__m128 __A) _mm512_broadcast_f32x4 (__m128 __A)
{ {
return (__m512) __builtin_ia32_broadcastf32x4_512 ((__v4sf) __A, return (__m512) __builtin_ia32_broadcastf32x4_512 ((__v4sf) __A,
(__v16sf) (__v16sf)
▲ Show 20 Lines • Show All 2,254 Lines • Show Last 20 Lines