diff --git a/src/pystencils/include/simd_horizontal_helpers.h b/src/pystencils/include/simd_horizontal_helpers.h index cd4bd5730657930f7a2375f74df0f5691ebaaf15..bd1889153a56876032781b77dcb66a413b3fe07d 100644 --- a/src/pystencils/include/simd_horizontal_helpers.h +++ b/src/pystencils/include/simd_horizontal_helpers.h @@ -120,7 +120,7 @@ inline float _mm256_horizontal_max_ps(float dst, __m256 src) { #endif -#if defined(__AVX512VL__) +#if defined(__AVX512F__) #include <immintrin.h> inline double _mm512_horizontal_add_pd(double dst, __m512d src) { diff --git a/util/generate_simd_horizontal_op.py b/util/generate_simd_horizontal_op.py index aebbf35bb65dd9fd66374a98a038ed3c4a3151db..1d652c6e1f53f5a7b060f2a725162ef07f1fdfae 100644 --- a/util/generate_simd_horizontal_op.py +++ b/util/generate_simd_horizontal_op.py @@ -277,7 +277,7 @@ vtypes_for_instruction_set = { guards_for_instruction_sets = { InstructionSets.SSE3: "__SSE3__", InstructionSets.AVX: "__AVX__", - InstructionSets.AVX512: '__AVX512VL__', + InstructionSets.AVX512: '__AVX512F__', InstructionSets.NEON: '_M_ARM64', }