lib/Headers/avxneconvertintrin.h

29 /// stored at memory locations starting at location \a __A to a
37 /// _mm_bcstnebf16_ps(const void *__A);
42 /// \param __A
49 /// b := Convert_BF16_To_FP32(MEM[__A+15:__A])
57 _mm_bcstnebf16_ps(const void *__A) {
58   return (__m128)__builtin_ia32_vbcstnebf162ps128((const __bf16 *)__A);
62 /// stored at memory locations starting at location \a __A to a
70 /// _mm256_bcstnebf16_ps(const void *__A);
75 /// \param __A
82 /// b := Convert_BF16_To_FP32(MEM[__A+15:__A])
90 _mm256_bcstnebf16_ps(const void *__A) {
91   return (__m256)__builtin_ia32_vbcstnebf162ps256((const __bf16 *)__A);
95 /// stored at memory locations starting at location \a __A to a
103 /// _mm_bcstnesh_ps(const void *__A);
108 /// \param __A
115 /// b := Convert_FP16_To_FP32(MEM[__A+15:__A])
123 _mm_bcstnesh_ps(const void *__A) {
124   return (__m128)__builtin_ia32_vbcstnesh2ps128((const _Float16 *)__A);
128 /// stored at memory locations starting at location \a __A to a
136 /// _mm256_bcstnesh_ps(const void *__A);
141 /// \param __A
148 /// b := Convert_FP16_To_FP32(MEM[__A+15:__A])
156 _mm256_bcstnesh_ps(const void *__A) {
157   return (__m256)__builtin_ia32_vbcstnesh2ps256((const _Float16 *)__A);
161 /// stored at memory locations starting at location \a __A to packed
168 /// _mm_cvtneebf16_ps(const __m128bh *__A);
173 /// \param __A
184 /// 	dst[m+31:m] := Convert_BF16_To_FP32(MEM[__A+i+15:__A+i])
189 _mm_cvtneebf16_ps(const __m128bh *__A) {
190   return (__m128)__builtin_ia32_vcvtneebf162ps128((const __v8bf *)__A);
194 /// stored at memory locations starting at location \a __A to packed
201 /// _mm256_cvtneebf16_ps(const __m256bh *__A);
206 /// \param __A
217 /// 	dst[m+31:m] := Convert_BF16_To_FP32(MEM[__A+i+15:__A+i])
222 _mm256_cvtneebf16_ps(const __m256bh *__A) {
223   return (__m256)__builtin_ia32_vcvtneebf162ps256((const __v16bf *)__A);
227 /// stored at memory locations starting at location \a __A to packed
234 /// _mm_cvtneeph_ps(const __m128h *__A);
239 /// \param __A
250 /// 	dst[m+31:m] := Convert_FP16_To_FP32(MEM[__A+i+15:__A+i])
255 _mm_cvtneeph_ps(const __m128h *__A) {
256   return (__m128)__builtin_ia32_vcvtneeph2ps128((const __v8hf *)__A);
260 /// stored at memory locations starting at location \a __A to packed
267 /// _mm256_cvtneeph_ps(const __m256h *__A);
272 /// \param __A
283 /// 	dst[m+31:m] := Convert_FP16_To_FP32(MEM[__A+i+15:__A+i])
288 _mm256_cvtneeph_ps(const __m256h *__A) {
289   return (__m256)__builtin_ia32_vcvtneeph2ps256((const __v16hf *)__A);
293 /// stored at memory locations starting at location \a __A to packed
300 /// _mm_cvtneobf16_ps(const __m128bh *__A);
305 /// \param __A
316 /// 	dst[m+31:m] := Convert_BF16_To_FP32(MEM[__A+i+15:__A+i])
321 _mm_cvtneobf16_ps(const __m128bh *__A) {
322   return (__m128)__builtin_ia32_vcvtneobf162ps128((const __v8bf *)__A);
326 /// stored at memory locations starting at location \a __A to packed
333 /// _mm256_cvtneobf16_ps(const __m256bh *__A);
338 /// \param __A
349 /// 	dst[m+31:m] := Convert_BF16_To_FP32(MEM[__A+i+15:__A+i])
354 _mm256_cvtneobf16_ps(const __m256bh *__A) {
355   return (__m256)__builtin_ia32_vcvtneobf162ps256((const __v16bf *)__A);
359 /// stored at memory locations starting at location \a __A to packed
366 /// _mm_cvtneoph_ps(const __m128h *__A);
371 /// \param __A
382 /// 	dst[m+31:m] := Convert_FP16_To_FP32(MEM[__A+i+15:__A+i])
387 _mm_cvtneoph_ps(const __m128h *__A) {
388   return (__m128)__builtin_ia32_vcvtneoph2ps128((const __v8hf *)__A);
392 /// stored at memory locations starting at location \a __A to packed
399 /// _mm256_cvtneoph_ps(const __m256h *__A);
404 /// \param __A
415 /// 	dst[m+31:m] := Convert_FP16_To_FP32(MEM[__A+i+15:__A+i])
420 _mm256_cvtneoph_ps(const __m256h *__A) {
421   return (__m256)__builtin_ia32_vcvtneoph2ps256((const __v16hf *)__A);
424 /// Convert packed single-precision (32-bit) floating-point elements in \a __A
431 /// _mm_cvtneps_avx_pbh(__m128 __A);
436 /// \param __A
443 /// 	dst.word[j] := Convert_FP32_To_BF16(__A.fp32[j])
448 _mm_cvtneps_avx_pbh(__m128 __A) {
449   return (__m128bh)__builtin_ia32_vcvtneps2bf16128((__v4sf)__A);
452 /// Convert packed single-precision (32-bit) floating-point elements in \a __A
459 /// _mm256_cvtneps_avx_pbh(__m256 __A);
464 /// \param __A
476 _mm256_cvtneps_avx_pbh(__m256 __A) {
477   return (__m128bh)__builtin_ia32_vcvtneps2bf16256((__v8sf)__A);