[clang] [llvm] [X86][AVX10.2] Support YMM rounding new instructions (PR #101825)
via llvm-commits
llvm-commits at lists.llvm.org
Sat Aug 3 07:10:07 PDT 2024
github-actions[bot] wrote:
<!--LLVM CODE FORMAT COMMENT: {clang-format}-->
:warning: C/C++ code formatter, clang-format found issues in your code. :warning:
<details>
<summary>
You can test this locally with the following command:
</summary>
``````````bash
git-clang-format --diff 79f7630e28589364ccf989a4a838f5dd74ce260a 520379066196ffc922c7d571399052a2eb7f869c --extensions cpp,c,h -- clang/lib/CodeGen/CGBuiltin.cpp clang/lib/Headers/avx10_2niintrin.h clang/lib/Sema/SemaX86.cpp clang/test/CodeGen/X86/avx10_2ni-builtins.c llvm/lib/Target/X86/Disassembler/X86Disassembler.cpp llvm/lib/Target/X86/X86InstrFMA3Info.cpp llvm/lib/Target/X86/X86IntrinsicsInfo.h llvm/utils/TableGen/X86DisassemblerTables.cpp
``````````
</details>
<details>
<summary>
View the diff from clang-format here.
</summary>
``````````diff
diff --git a/clang/lib/Headers/avx10_2niintrin.h b/clang/lib/Headers/avx10_2niintrin.h
index 42b24d2b5b..830028650f 100644
--- a/clang/lib/Headers/avx10_2niintrin.h
+++ b/clang/lib/Headers/avx10_2niintrin.h
@@ -81,7 +81,7 @@
#define _mm256_cmp_round_pd_mask(A, B, P, R) \
((__mmask8)__builtin_ia32_vcmppd256_round_mask( \
- (__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (int)(P), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (int)(P), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cmp_round_pd_mask(U, A, B, P, R) \
@@ -91,7 +91,7 @@
#define _mm256_cmp_round_ph_mask(A, B, P, R) \
((__mmask16)__builtin_ia32_vcmpph256_round_mask( \
- (__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), (int)(P), (__mmask16)-1, \
+ (__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), (int)(P), (__mmask16) - 1, \
(int)(R)))
#define _mm256_mask_cmp_round_ph_mask(U, A, B, P, R) \
@@ -101,7 +101,7 @@
#define _mm256_cmp_round_ps_mask(A, B, P, R) \
((__mmask8)__builtin_ia32_vcmpps256_round_mask( \
- (__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (int)(P), (__mmask8)-1, \
+ (__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (int)(P), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cmp_round_ps_mask(U, A, B, P, R) \
@@ -124,7 +124,7 @@
#define _mm256_cvt_roundepi32_ps(A, R) \
((__m256)__builtin_ia32_vcvtdq2ps256_round_mask((__v8si)(__m256i)(A), \
(__v8sf)_mm256_setzero_ps(), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_cvt_roundepi32_ps(W, U, A, R) \
((__m256)__builtin_ia32_vcvtdq2ps256_round_mask( \
@@ -137,7 +137,7 @@
#define _mm256_cvt_roundpd_epi32(A, R) \
((__m128i)__builtin_ia32_vcvtpd2dq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4si)_mm_setzero_si128(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4si)_mm_setzero_si128(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundpd_epi32(W, U, A, R) \
@@ -162,8 +162,9 @@
(__v4df)(A), (__v8hf)_mm_setzero_ph(), (__mmask8)(U), (int)(R)))
#define _mm256_cvt_roundpd_ps(A, R) \
- ((__m128)__builtin_ia32_vcvtpd2ps256_round_mask( \
- (__v4df)(__m256d)(A), (__v4sf)_mm_setzero_ps(), (__mmask8)-1, (int)(R)))
+ ((__m128)__builtin_ia32_vcvtpd2ps256_round_mask((__v4df)(__m256d)(A), \
+ (__v4sf)_mm_setzero_ps(), \
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_cvt_roundpd_ps(W, U, A, R) \
((__m128)__builtin_ia32_vcvtpd2ps256_round_mask( \
@@ -176,7 +177,7 @@
#define _mm256_cvt_roundpd_epi64(A, R) \
((__m256i)__builtin_ia32_vcvtpd2qq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4di)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4di)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundpd_epi64(W, U, A, R) \
@@ -190,7 +191,7 @@
#define _mm256_cvt_roundpd_epu32(A, R) \
((__m128i)__builtin_ia32_vcvtpd2udq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4su)_mm_setzero_si128(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4su)_mm_setzero_si128(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundpd_epu32(W, U, A, R) \
@@ -204,7 +205,7 @@
#define _mm256_cvt_roundpd_epu64(A, R) \
((__m256i)__builtin_ia32_vcvtpd2uqq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4du)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4du)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundpd_epu64(W, U, A, R) \
@@ -322,7 +323,7 @@
#define _mm256_cvt_roundps_epi32(A, R) \
((__m256i)__builtin_ia32_vcvtps2dq256_round_mask( \
- (__v8sf)(__m256)(A), (__v8si)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v8sf)(__m256)(A), (__v8si)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundps_epi32(W, U, A, R) \
@@ -336,7 +337,7 @@
#define _mm256_cvt_roundps_pd(A, R) \
((__m256d)__builtin_ia32_vcvtps2pd256_round_mask( \
- (__v4sf)(__m128)(A), (__v4df)_mm256_undefined_pd(), (__mmask8)-1, \
+ (__v4sf)(__m128)(A), (__v4df)_mm256_undefined_pd(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundps_pd(W, U, A, R) \
@@ -351,7 +352,7 @@
#define _mm256_cvt_roundps_ph(A, I) \
((__m128i)__builtin_ia32_vcvtps2ph256_mask((__v8sf)(__m256)(A), (int)(I), \
(__v8hi)_mm_undefined_si128(), \
- (__mmask8)-1))
+ (__mmask8) - 1))
/* FIXME: We may use these way in future.
#define _mm256_cvt_roundps_ph(A, I) \
@@ -380,7 +381,7 @@
#define _mm256_cvt_roundps_epi64(A, R) \
((__m256i)__builtin_ia32_vcvtps2qq256_round_mask( \
- (__v4sf)(__m128)(A), (__v4di)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4sf)(__m128)(A), (__v4di)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundps_epi64(W, U, A, R) \
@@ -394,7 +395,7 @@
#define _mm256_cvt_roundps_epu32(A, R) \
((__m256i)__builtin_ia32_vcvtps2udq256_round_mask( \
- (__v8sf)(__m256)(A), (__v8su)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v8sf)(__m256)(A), (__v8su)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundps_epu32(W, U, A, R) \
@@ -408,7 +409,7 @@
#define _mm256_cvt_roundps_epu64(A, R) \
((__m256i)__builtin_ia32_vcvtps2uqq256_round_mask( \
- (__v4sf)(__m128)(A), (__v4du)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4sf)(__m128)(A), (__v4du)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundps_epu64(W, U, A, R) \
@@ -422,7 +423,7 @@
#define _mm256_cvt_roundepi64_pd(A, R) \
((__m256d)__builtin_ia32_vcvtqq2pd256_round_mask( \
- (__v4di)(__m256i)(A), (__v4df)_mm256_setzero_pd(), (__mmask8)-1, \
+ (__v4di)(__m256i)(A), (__v4df)_mm256_setzero_pd(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundepi64_pd(W, U, A, R) \
@@ -447,8 +448,9 @@
(__v4di)(A), (__v8hf)_mm_setzero_ph(), (__mmask8)(U), (int)(R)))
#define _mm256_cvt_roundepi64_ps(A, R) \
- ((__m128)__builtin_ia32_vcvtqq2ps256_round_mask( \
- (__v4di)(__m256i)(A), (__v4sf)_mm_setzero_ps(), (__mmask8)-1, (int)(R)))
+ ((__m128)__builtin_ia32_vcvtqq2ps256_round_mask((__v4di)(__m256i)(A), \
+ (__v4sf)_mm_setzero_ps(), \
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_cvt_roundepi64_ps(W, U, A, R) \
((__m128)__builtin_ia32_vcvtqq2ps256_round_mask( \
@@ -461,7 +463,7 @@
#define _mm256_cvtt_roundpd_epi32(A, R) \
((__m128i)__builtin_ia32_vcvttpd2dq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4si)_mm_setzero_si128(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4si)_mm_setzero_si128(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundpd_epi32(W, U, A, R) \
@@ -475,7 +477,7 @@
#define _mm256_cvtt_roundpd_epi64(A, R) \
((__m256i)__builtin_ia32_vcvttpd2qq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4di)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4di)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundpd_epi64(W, U, A, R) \
@@ -489,7 +491,7 @@
#define _mm256_cvtt_roundpd_epu32(A, R) \
((__m128i)__builtin_ia32_vcvttpd2udq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4su)_mm_setzero_si128(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4su)_mm_setzero_si128(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundpd_epu32(W, U, A, R) \
@@ -503,7 +505,7 @@
#define _mm256_cvtt_roundpd_epu64(A, R) \
((__m256i)__builtin_ia32_vcvttpd2uqq256_round_mask( \
- (__v4df)(__m256d)(A), (__v4du)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4du)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundpd_epu64(W, U, A, R) \
@@ -597,7 +599,7 @@
#define _mm256_cvtt_roundps_epi32(A, R) \
((__m256i)__builtin_ia32_vcvttps2dq256_round_mask( \
- (__v8sf)(__m256)(A), (__v8si)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v8sf)(__m256)(A), (__v8si)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundps_epi32(W, U, A, R) \
@@ -611,7 +613,7 @@
#define _mm256_cvtt_roundps_epi64(A, R) \
((__m256i)__builtin_ia32_vcvttps2qq256_round_mask( \
- (__v4sf)(__m128)(A), (__v4di)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4sf)(__m128)(A), (__v4di)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundps_epi64(W, U, A, R) \
@@ -625,7 +627,7 @@
#define _mm256_cvtt_roundps_epu32(A, R) \
((__m256i)__builtin_ia32_vcvttps2udq256_round_mask( \
- (__v8sf)(__m256)(A), (__v8su)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v8sf)(__m256)(A), (__v8su)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundps_epu32(W, U, A, R) \
@@ -639,7 +641,7 @@
#define _mm256_cvtt_roundps_epu64(A, R) \
((__m256i)__builtin_ia32_vcvttps2uqq256_round_mask( \
- (__v4sf)(__m128)(A), (__v4du)_mm256_setzero_si256(), (__mmask8)-1, \
+ (__v4sf)(__m128)(A), (__v4du)_mm256_setzero_si256(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvtt_roundps_epu64(W, U, A, R) \
@@ -665,7 +667,7 @@
#define _mm256_cvt_roundepu32_ps(A, R) \
((__m256)__builtin_ia32_vcvtudq2ps256_round_mask( \
- (__v8su)(__m256i)(A), (__v8sf)_mm256_setzero_ps(), (__mmask8)-1, \
+ (__v8su)(__m256i)(A), (__v8sf)_mm256_setzero_ps(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundepu32_ps(W, U, A, R) \
@@ -679,7 +681,7 @@
#define _mm256_cvt_roundepu64_pd(A, R) \
((__m256d)__builtin_ia32_vcvtuqq2pd256_round_mask( \
- (__v4du)(__m256i)(A), (__v4df)_mm256_setzero_pd(), (__mmask8)-1, \
+ (__v4du)(__m256i)(A), (__v4df)_mm256_setzero_pd(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_cvt_roundepu64_pd(W, U, A, R) \
@@ -704,8 +706,9 @@
(__v4du)(A), (__v8hf)_mm_setzero_ph(), (__mmask8)(U), (int)(R)))
#define _mm256_cvt_roundepu64_ps(A, R) \
- ((__m128)__builtin_ia32_vcvtuqq2ps256_round_mask( \
- (__v4du)(__m256i)(A), (__v4sf)_mm_setzero_ps(), (__mmask8)-1, (int)(R)))
+ ((__m128)__builtin_ia32_vcvtuqq2ps256_round_mask((__v4du)(__m256i)(A), \
+ (__v4sf)_mm_setzero_ps(), \
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_cvt_roundepu64_ps(W, U, A, R) \
((__m128)__builtin_ia32_vcvtuqq2ps256_round_mask( \
@@ -787,7 +790,7 @@
#define _mm256_fcmadd_round_pch(A, B, C, R) \
((__m256h)__builtin_ia32_vfcmaddcph256_round_mask3( \
(__v8sf)(__m256h)(A), (__v8sf)(__m256h)(B), (__v8sf)(__m256h)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fcmadd_round_pch(A, U, B, C, R) \
((__m256h)__builtin_ia32_vfcmaddcph256_round_mask( \
@@ -807,7 +810,7 @@
#define _mm256_cmul_round_pch(A, B, R) \
((__m256h)__builtin_ia32_vfcmulcph256_round_mask( \
(__v8sf)(__m256h)(A), (__v8sf)(__m256h)(B), \
- (__v8sf)(__m256h)_mm256_undefined_ph(), (__mmask8)-1, (int)(R)))
+ (__v8sf)(__m256h)_mm256_undefined_ph(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_cmul_round_pch(W, U, A, B, R) \
((__m256h)__builtin_ia32_vfcmulcph256_round_mask( \
@@ -822,7 +825,7 @@
#define _mm256_fixupimm_round_pd(A, B, C, imm, R) \
((__m256d)__builtin_ia32_vfixupimmpd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (__v4di)(__m256i)(C), \
- (int)(imm), (__mmask8)-1, (int)(R)))
+ (int)(imm), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fixupimm_round_pd(A, U, B, C, imm, R) \
((__m256d)__builtin_ia32_vfixupimmpd256_round_mask( \
@@ -837,7 +840,7 @@
#define _mm256_fixupimm_round_ps(A, B, C, imm, R) \
((__m256)__builtin_ia32_vfixupimmps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (__v8si)(__m256i)(C), \
- (int)(imm), (__mmask8)-1, (int)(R)))
+ (int)(imm), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fixupimm_round_ps(A, U, B, C, imm, R) \
((__m256)__builtin_ia32_vfixupimmps256_round_mask( \
@@ -852,7 +855,7 @@
#define _mm256_fmadd_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (__v4df)(__m256d)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmadd_round_pd(A, U, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask( \
@@ -872,7 +875,7 @@
#define _mm256_fmsub_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), -(__v4df)(__m256d)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmsub_round_pd(A, U, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask( \
@@ -887,7 +890,7 @@
#define _mm256_fnmadd_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask( \
-(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (__v4df)(__m256d)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask3_fnmadd_round_pd(A, B, C, U, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask3( \
@@ -902,7 +905,7 @@
#define _mm256_fnmsub_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_mask( \
-(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), -(__v4df)(__m256d)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_maskz_fnmsub_round_pd(U, A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddpd256_round_maskz( \
@@ -912,7 +915,7 @@
#define _mm256_fmadd_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask( \
(__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), (__v16hf)(__m256h)(C), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_mask_fmadd_round_ph(A, U, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask( \
@@ -932,7 +935,7 @@
#define _mm256_fmsub_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask( \
(__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), -(__v16hf)(__m256h)(C), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_mask_fmsub_round_ph(A, U, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask( \
@@ -947,7 +950,7 @@
#define _mm256_fnmadd_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask( \
(__v16hf)(__m256h)(A), -(__v16hf)(__m256h)(B), (__v16hf)(__m256h)(C), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_mask3_fnmadd_round_ph(A, B, C, U, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask3( \
@@ -962,7 +965,7 @@
#define _mm256_fnmsub_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_mask( \
(__v16hf)(__m256h)(A), -(__v16hf)(__m256h)(B), -(__v16hf)(__m256h)(C), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_maskz_fnmsub_round_ph(U, A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddph256_round_maskz( \
@@ -972,7 +975,7 @@
#define _mm256_fmadd_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (__v8sf)(__m256)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmadd_round_ps(A, U, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask( \
@@ -992,7 +995,7 @@
#define _mm256_fmsub_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), -(__v8sf)(__m256)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmsub_round_ps(A, U, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask( \
@@ -1007,7 +1010,7 @@
#define _mm256_fnmadd_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask( \
(__v8sf)(__m256)(A), -(__v8sf)(__m256)(B), (__v8sf)(__m256)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask3_fnmadd_round_ps(A, B, C, U, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask3( \
@@ -1022,7 +1025,7 @@
#define _mm256_fnmsub_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_mask( \
(__v8sf)(__m256)(A), -(__v8sf)(__m256)(B), -(__v8sf)(__m256)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_maskz_fnmsub_round_ps(U, A, B, C, R) \
((__m256)__builtin_ia32_vfmaddps256_round_maskz( \
@@ -1032,7 +1035,7 @@
#define _mm256_fmadd_round_pch(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddcph256_round_mask3( \
(__v8sf)(__m256h)(A), (__v8sf)(__m256h)(B), (__v8sf)(__m256h)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmadd_round_pch(A, U, B, C, R) \
((__m256h)__builtin_ia32_vfmaddcph256_round_mask( \
@@ -1052,7 +1055,7 @@
#define _mm256_fmaddsub_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddsubpd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (__v4df)(__m256d)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmaddsub_round_pd(A, U, B, C, R) \
((__m256d)__builtin_ia32_vfmaddsubpd256_round_mask( \
@@ -1072,7 +1075,7 @@
#define _mm256_fmsubadd_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vfmaddsubpd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), -(__v4df)(__m256d)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmsubadd_round_pd(A, U, B, C, R) \
((__m256d)__builtin_ia32_vfmaddsubpd256_round_mask( \
@@ -1087,7 +1090,7 @@
#define _mm256_fmaddsub_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddsubph256_round_mask( \
(__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), (__v16hf)(__m256h)(C), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_mask_fmaddsub_round_ph(A, U, B, C, R) \
((__m256h)__builtin_ia32_vfmaddsubph256_round_mask( \
@@ -1107,7 +1110,7 @@
#define _mm256_fmsubadd_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vfmaddsubph256_round_mask( \
(__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), -(__v16hf)(__m256h)(C), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_mask_fmsubadd_round_ph(A, U, B, C, R) \
((__m256h)__builtin_ia32_vfmaddsubph256_round_mask( \
@@ -1122,7 +1125,7 @@
#define _mm256_fmaddsub_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vfmaddsubps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (__v8sf)(__m256)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmaddsub_round_ps(A, U, B, C, R) \
((__m256)__builtin_ia32_vfmaddsubps256_round_mask( \
@@ -1142,7 +1145,7 @@
#define _mm256_fmsubadd_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vfmaddsubps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), -(__v8sf)(__m256)(C), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_fmsubadd_round_ps(A, U, B, C, R) \
((__m256)__builtin_ia32_vfmaddsubps256_round_mask( \
@@ -1231,7 +1234,7 @@
#define _mm256_mul_round_pch(A, B, R) \
((__m256h)__builtin_ia32_vfmulcph256_round_mask( \
(__v8sf)(__m256h)(A), (__v8sf)(__m256h)(B), \
- (__v8sf)(__m256h)_mm256_undefined_ph(), (__mmask8)-1, (int)(R)))
+ (__v8sf)(__m256h)_mm256_undefined_ph(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_mul_round_pch(W, U, A, B, R) \
((__m256h)__builtin_ia32_vfmulcph256_round_mask( \
@@ -1245,7 +1248,7 @@
#define _mm256_getexp_round_pd(A, R) \
((__m256d)__builtin_ia32_vgetexppd256_round_mask( \
- (__v4df)(__m256d)(A), (__v4df)_mm256_undefined_pd(), (__mmask8)-1, \
+ (__v4df)(__m256d)(A), (__v4df)_mm256_undefined_pd(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_getexp_round_pd(W, U, A, R) \
@@ -1259,7 +1262,7 @@
#define _mm256_getexp_round_ph(A, R) \
((__m256h)__builtin_ia32_vgetexpph256_round_mask( \
- (__v16hf)(__m256h)(A), (__v16hf)_mm256_undefined_ph(), (__mmask16)-1, \
+ (__v16hf)(__m256h)(A), (__v16hf)_mm256_undefined_ph(), (__mmask16) - 1, \
(int)(R)))
#define _mm256_mask_getexp_round_ph(W, U, A, R) \
@@ -1273,7 +1276,7 @@
#define _mm256_getexp_round_ps(A, R) \
((__m256)__builtin_ia32_vgetexpps256_round_mask( \
- (__v8sf)(__m256)(A), (__v8sf)_mm256_undefined_ps(), (__mmask8)-1, \
+ (__v8sf)(__m256)(A), (__v8sf)_mm256_undefined_ps(), (__mmask8) - 1, \
(int)(R)))
#define _mm256_mask_getexp_round_ps(W, U, A, R) \
@@ -1288,7 +1291,7 @@
#define _mm256_getmant_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vgetmantpd256_round_mask( \
(__v4df)(__m256d)(A), (int)(((C) << 2) | (B)), \
- (__v4df)_mm256_undefined_pd(), (__mmask8)-1, (int)(R)))
+ (__v4df)_mm256_undefined_pd(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_getmant_round_pd(W, U, A, B, C, R) \
((__m256d)__builtin_ia32_vgetmantpd256_round_mask( \
@@ -1303,7 +1306,7 @@
#define _mm256_getmant_round_ph(A, B, C, R) \
((__m256h)__builtin_ia32_vgetmantph256_round_mask( \
(__v16hf)(__m256h)(A), (int)(((C) << 2) | (B)), \
- (__v16hf)_mm256_undefined_ph(), (__mmask16)-1, (int)(R)))
+ (__v16hf)_mm256_undefined_ph(), (__mmask16) - 1, (int)(R)))
#define _mm256_mask_getmant_round_ph(W, U, A, B, C, R) \
((__m256h)__builtin_ia32_vgetmantph256_round_mask( \
@@ -1318,7 +1321,7 @@
#define _mm256_getmant_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vgetmantps256_round_mask( \
(__v8sf)(__m256)(A), (int)(((C) << 2) | (B)), \
- (__v8sf)_mm256_undefined_ps(), (__mmask8)-1, (int)(R)))
+ (__v8sf)_mm256_undefined_ps(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_getmant_round_ps(W, U, A, B, C, R) \
((__m256)__builtin_ia32_vgetmantps256_round_mask( \
@@ -1459,7 +1462,7 @@
#define _mm256_range_round_pd(A, B, C, R) \
((__m256d)__builtin_ia32_vrangepd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), (int)(C), \
- (__v4df)_mm256_setzero_pd(), (__mmask8)-1, (int)(R)))
+ (__v4df)_mm256_setzero_pd(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_range_round_pd(W, U, A, B, C, R) \
((__m256d)__builtin_ia32_vrangepd256_round_mask( \
@@ -1474,7 +1477,7 @@
#define _mm256_range_round_ps(A, B, C, R) \
((__m256)__builtin_ia32_vrangeps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (int)(C), \
- (__v8sf)_mm256_setzero_ps(), (__mmask8)-1, (int)(R)))
+ (__v8sf)_mm256_setzero_ps(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_range_round_ps(W, U, A, B, C, R) \
((__m256)__builtin_ia32_vrangeps256_round_mask( \
@@ -1489,7 +1492,7 @@
#define _mm256_reduce_round_pd(A, B, R) \
((__m256d)__builtin_ia32_vreducepd256_round_mask( \
(__v4df)(__m256d)(A), (int)(B), (__v4df)_mm256_setzero_pd(), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_reduce_round_pd(W, U, A, B, R) \
((__m256d)__builtin_ia32_vreducepd256_round_mask( \
@@ -1514,12 +1517,12 @@
#define _mm256_reduce_round_ph(A, imm, R) \
((__m256h)__builtin_ia32_vreduceph256_round_mask( \
(__v16hf)(__m256h)(A), (int)(imm), (__v16hf)_mm256_undefined_ph(), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_reduce_round_ps(A, B, R) \
((__m256)__builtin_ia32_vreduceps256_round_mask( \
(__v8sf)(__m256)(A), (int)(B), (__v8sf)_mm256_setzero_ps(), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_reduce_round_ps(W, U, A, B, R) \
((__m256)__builtin_ia32_vreduceps256_round_mask( \
@@ -1534,7 +1537,7 @@
#define _mm256_roundscale_round_pd(A, imm, R) \
((__m256d)__builtin_ia32_vrndscalepd256_round_mask( \
(__v4df)(__m256d)(A), (int)(imm), (__v4df)_mm256_undefined_pd(), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_roundscale_round_pd(A, B, C, imm, R) \
((__m256d)__builtin_ia32_vrndscalepd256_round_mask( \
@@ -1549,7 +1552,7 @@
#define _mm256_roundscale_round_ph(A, imm, R) \
((__m256h)__builtin_ia32_vrndscaleph256_round_mask( \
(__v16hf)(__m256h)(A), (int)(imm), (__v16hf)_mm256_undefined_ph(), \
- (__mmask16)-1, (int)(R)))
+ (__mmask16) - 1, (int)(R)))
#define _mm256_mask_roundscale_round_ph(A, B, C, imm, R) \
((__m256h)__builtin_ia32_vrndscaleph256_round_mask( \
@@ -1564,7 +1567,7 @@
#define _mm256_roundscale_round_ps(A, imm, R) \
((__m256)__builtin_ia32_vrndscaleps256_round_mask( \
(__v8sf)(__m256)(A), (int)(imm), (__v8sf)_mm256_undefined_ps(), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_roundscale_round_ps(A, B, C, imm, R) \
((__m256)__builtin_ia32_vrndscaleps256_round_mask( \
@@ -1579,7 +1582,7 @@
#define _mm256_scalef_round_pd(A, B, R) \
((__m256d)__builtin_ia32_vscalefpd256_round_mask( \
(__v4df)(__m256d)(A), (__v4df)(__m256d)(B), \
- (__v4df)_mm256_undefined_pd(), (__mmask8)-1, (int)(R)))
+ (__v4df)_mm256_undefined_pd(), (__mmask8) - 1, (int)(R)))
#define _mm256_mask_scalef_round_pd(W, U, A, B, R) \
((__m256d)__builtin_ia32_vscalefpd256_round_mask( \
@@ -1594,7 +1597,7 @@
#define _mm256_scalef_round_ph(A, B, R) \
((__m256h)__builtin_ia32_vscalefph256_round_mask( \
(__v16hf)(__m256h)(A), (__v16hf)(__m256h)(B), \
- (__v16hf)_mm256_undefined_ph(), (__mmask16)-1, (int)(R)))
+ (__v16hf)_mm256_undefined_ph(), (__mmask16) - 1, (int)(R)))
#define _mm256_mask_scalef_round_ph(W, U, A, B, R) \
((__m256h)__builtin_ia32_vscalefph256_round_mask( \
@@ -1609,7 +1612,7 @@
#define _mm256_scalef_round_ps(A, B, R) \
((__m256)__builtin_ia32_vscalefps256_round_mask( \
(__v8sf)(__m256)(A), (__v8sf)(__m256)(B), (__v8sf)_mm256_undefined_ps(), \
- (__mmask8)-1, (int)(R)))
+ (__mmask8) - 1, (int)(R)))
#define _mm256_mask_scalef_round_ps(W, U, A, B, R) \
((__m256)__builtin_ia32_vscalefps256_round_mask( \
``````````
</details>
https://github.com/llvm/llvm-project/pull/101825
More information about the llvm-commits
mailing list