This is an archive of the discontinued LLVM Phabricator instance.

Differential D21734

[Intrinsics][AVX512][BuiltIn] adding intrinsics for vrangesd instruction set
ClosedPublic

Authored by m_zuckerman on Jun 26 2016, 6:24 AM.

Details

Reviewers
AsafBadouh
delena
igorb
Commits
rG3f316abdce3a: [Clang][Intrinsics][AVX512][BuiltIn] adding intrinsics for vrangesd instruction…
rC274218: [Clang][Intrinsics][AVX512][BuiltIn] adding intrinsics for vrangesd instruction…
rL274218: [Clang][Intrinsics][AVX512][BuiltIn] adding intrinsics for vrangesd…

Diff Detail

Event Timeline

m_zuckerman updated this revision to Diff 61907.Jun 26 2016, 6:24 AM
m_zuckerman retitled this revision from to [Intrinsics][AVX512][BuiltIn] adding intrinsics for vrangesd instruction set.
m_zuckerman updated this object.
m_zuckerman added reviewers: AsafBadouh, igorb, delena.
m_zuckerman added a subscriber: cfe-commits.
igorb accepted this revision.Jun 26 2016, 1:29 PM
igorb edited edge metadata.

LGTM

This revision is now accepted and ready to land.Jun 26 2016, 1:29 PM
Closed by commit rL274218: [Clang][Intrinsics][AVX512][BuiltIn] adding intrinsics for vrangesd… (authored by mzuckerm). · Explain WhyJun 30 2016, 1:12 AM
This revision was automatically updated to reflect the committed changes.

Revision Contents

PathSize
lib/
Headers/
12 lines
test/
CodeGen/
36 lines

Diff 61907

lib/Headers/avx512dqintrin.h

Show First 20 Lines • Show All 786 Lines • ▼ Show 20 Lines
#define _mm_range_round_ss(A, B, C, R) __extension__ ({ \ #define _mm_range_round_ss(A, B, C, R) __extension__ ({ \
(__m128)__builtin_ia32_rangess128_round_mask((__v4sf)(__m128)(A), \ (__m128)__builtin_ia32_rangess128_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \ (__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \ (__v4sf)_mm_setzero_ps(), \
(__mmask8) -1, (int)(C),\ (__mmask8) -1, (int)(C),\
(int)(R)); }) (int)(R)); })
#define _mm_range_ss(A ,B , C) _mm_range_round_ss(A, B, C ,_MM_FROUND_CUR_DIRECTION)
#define _mm_mask_range_round_ss(W, U, A, B, C, R) __extension__ ({ \ #define _mm_mask_range_round_ss(W, U, A, B, C, R) __extension__ ({ \
(__m128)__builtin_ia32_rangess128_round_mask((__v4sf)(__m128)(A), \ (__m128)__builtin_ia32_rangess128_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \ (__v4sf)(__m128)(B), \
(__v4sf)(__m128)(W),\ (__v4sf)(__m128)(W),\
(__mmask8)(U), (int)(C),\ (__mmask8)(U), (int)(C),\
(int)(R)); }) (int)(R)); })
#define _mm_mask_range_ss(W , U, A, B, C) _mm_mask_range_round_ss(W, U, A, B, C , _MM_FROUND_CUR_DIRECTION)
#define _mm_maskz_range_round_ss(U, A, B, C, R) __extension__ ({ \ #define _mm_maskz_range_round_ss(U, A, B, C, R) __extension__ ({ \
(__m128)__builtin_ia32_rangess128_round_mask((__v4sf)(__m128)(A), \ (__m128)__builtin_ia32_rangess128_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \ (__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \ (__v4sf)_mm_setzero_ps(), \
(__mmask8)(U), (int)(C),\ (__mmask8)(U), (int)(C),\
(int)(R)); }) (int)(R)); })
#define _mm_maskz_range_ss(U, A ,B , C) _mm_maskz_range_round_ss(U, A, B, C ,_MM_FROUND_CUR_DIRECTION)
#define _mm_range_round_sd(A, B, C, R) __extension__ ({ \ #define _mm_range_round_sd(A, B, C, R) __extension__ ({ \
(__m128d)__builtin_ia32_rangesd128_round_mask((__v2df)(__m128d)(A), \ (__m128d)__builtin_ia32_rangesd128_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \ (__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \ (__v2df)_mm_setzero_pd(), \
(__mmask8) -1, (int)(C),\ (__mmask8) -1, (int)(C),\
(int)(R)); }) (int)(R)); })
#define _mm_range_sd(A ,B , C) _mm_range_round_sd(A, B, C ,_MM_FROUND_CUR_DIRECTION)
#define _mm_mask_range_round_sd(W, U, A, B, C, R) __extension__ ({ \ #define _mm_mask_range_round_sd(W, U, A, B, C, R) __extension__ ({ \
(__m128d)__builtin_ia32_rangesd128_round_mask((__v2df)(__m128d)(A), \ (__m128d)__builtin_ia32_rangesd128_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \ (__v2df)(__m128d)(B), \
(__v2df)(__m128d)(W),\ (__v2df)(__m128d)(W),\
(__mmask8)(U), (int)(C),\ (__mmask8)(U), (int)(C),\
(int)(R)); }) (int)(R)); })
#define _mm_mask_range_sd(W, U, A, B, C) _mm_mask_range_round_sd(W, U, A, B, C ,_MM_FROUND_CUR_DIRECTION)
#define _mm_maskz_range_round_sd(U, A, B, C, R) __extension__ ({ \ #define _mm_maskz_range_round_sd(U, A, B, C, R) __extension__ ({ \
(__m128d)__builtin_ia32_rangesd128_round_mask((__v2df)(__m128d)(A), \ (__m128d)__builtin_ia32_rangesd128_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \ (__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \ (__v2df)_mm_setzero_pd(), \
(__mmask8)(U), (int)(C),\ (__mmask8)(U), (int)(C),\
(int)(R)); }) (int)(R)); })
#define _mm_maskz_range_sd(U, A, B, C) _mm_maskz_range_round_sd(U, A, B, C ,_MM_FROUND_CUR_DIRECTION)
#define _mm512_reduce_pd(A, B) __extension__ ({ \ #define _mm512_reduce_pd(A, B) __extension__ ({ \
(__m512d)__builtin_ia32_reducepd512_mask((__v8df)(__m512d)(A), (int)(B), \ (__m512d)__builtin_ia32_reducepd512_mask((__v8df)(__m512d)(A), (int)(B), \
(__v8df)_mm512_setzero_pd(), \ (__v8df)_mm512_setzero_pd(), \
(__mmask8)-1, \ (__mmask8)-1, \
_MM_FROUND_CUR_DIRECTION); }) _MM_FROUND_CUR_DIRECTION); })
#define _mm512_mask_reduce_pd(W, U, A, B) __extension__ ({ \ #define _mm512_mask_reduce_pd(W, U, A, B) __extension__ ({ \
(__m512d)__builtin_ia32_reducepd512_mask((__v8df)(__m512d)(A), (int)(B), \ (__m512d)__builtin_ia32_reducepd512_mask((__v8df)(__m512d)(A), (int)(B), \
▲ Show 20 Lines • Show All 482 Lines • Show Last 20 Lines

test/CodeGen/avx512dq-builtins.c

Show First 20 Lines • Show All 665 Lines • ▼ Show 20 Lines
} }
__m128 test_mm512_maskz_range_round_ss(__mmask8 __U, __m128 __A, __m128 __B) { __m128 test_mm512_maskz_range_round_ss(__mmask8 __U, __m128 __A, __m128 __B) {
// CHECK-LABEL: @test_mm512_maskz_range_round_ss // CHECK-LABEL: @test_mm512_maskz_range_round_ss
// CHECK: @llvm.x86.avx512.mask.range.ss // CHECK: @llvm.x86.avx512.mask.range.ss
return _mm_maskz_range_round_ss(__U, __A, __B, 4, 8); return _mm_maskz_range_round_ss(__U, __A, __B, 4, 8);
} }
__m128d test_mm_range_sd(__m128d __A, __m128d __B) {
// CHECK-LABEL: @test_mm_range_sd
// CHECK: @llvm.x86.avx512.mask.range.sd
return _mm_range_sd(__A, __B, 4);
}
__m128d test_mm_mask_range_sd(__m128d __W, __mmask8 __U, __m128d __A, __m128d __B) {
// CHECK-LABEL: test_mm_mask_range_sd
// CHECK: @llvm.x86.avx512.mask.range.sd
return _mm_mask_range_sd(__W, __U, __A, __B, 4);
}
__m128d test_mm_maskz_range_sd(__mmask8 __U, __m128d __A, __m128d __B) {
// CHECK-LABEL: @test_mm_maskz_range_sd
// CHECK: @llvm.x86.avx512.mask.range.sd
return _mm_maskz_range_sd(__U, __A, __B, 4);
}
__m128d test_mm_range_ss(__m128d __A, __m128d __B) {
// CHECK-LABEL: @test_mm_range_ss
// CHECK: @llvm.x86.avx512.mask.range.ss
return _mm_range_ss(__A, __B, 4);
}
__m128d test_mm_mask_range_ss(__m128 __W, __mmask8 __U, __m128 __A, __m128 __B) {
// CHECK-LABEL: @test_mm_mask_range_ss
// CHECK: @llvm.x86.avx512.mask.range.ss
return _mm_mask_range_ss(__W, __U, __A, __B, 4);
}
__m128 test_mm_maskz_range_ss(__mmask8 __U, __m128 __A, __m128 __B) {
// CHECK-LABEL: @test_mm_maskz_range_ss
// CHECK: @llvm.x86.avx512.mask.range.ss
return _mm_maskz_range_ss(__U, __A, __B, 4);
}
__m512 test_mm512_range_ps(__m512 __A, __m512 __B) { __m512 test_mm512_range_ps(__m512 __A, __m512 __B) {
// CHECK-LABEL: @test_mm512_range_ps // CHECK-LABEL: @test_mm512_range_ps
// CHECK: @llvm.x86.avx512.mask.range.ps.512 // CHECK: @llvm.x86.avx512.mask.range.ps.512
return _mm512_range_ps(__A, __B, 4); return _mm512_range_ps(__A, __B, 4);
} }
__m512 test_mm512_mask_range_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B) { __m512 test_mm512_mask_range_ps(__m512 __W, __mmask16 __U, __m512 __A, __m512 __B) {
// CHECK-LABEL: @test_mm512_mask_range_ps // CHECK-LABEL: @test_mm512_mask_range_ps
▲ Show 20 Lines • Show All 494 Lines • Show Last 20 Lines