(__mmask8) __U);
}
+extern __inline __m256i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_or_epi32 (__m256i __A, __m256i __B)
+{
+ return (__m256i) ((__v8su)__A | (__v8su)__B);
+}
+
extern __inline __m128i
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_or_epi32 (__m128i __W, __mmask8 __U, __m128i __A, __m128i __B)
(__mmask8) __U);
}
+extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_or_epi32 (__m128i __A, __m128i __B)
+{
+ return (__m128i) ((__v4su)__A | (__v4su)__B);
+}
+
extern __inline __m256i
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm256_mask_xor_epi32 (__m256i __W, __mmask8 __U, __m256i __A,
(__mmask8) __U);
}
+extern __inline __m256i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_xor_epi32 (__m256i __A, __m256i __B)
+{
+ return (__m256i) ((__v8su)__A ^ (__v8su)__B);
+}
+
extern __inline __m128i
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_xor_epi32 (__m128i __W, __mmask8 __U, __m128i __A,
(__mmask8) __U);
}
+extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_xor_epi32 (__m128i __A, __m128i __B)
+{
+ return (__m128i) ((__v4su)__A ^ (__v4su)__B);
+}
+
extern __inline __m128
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_cvtpd_ps (__m128 __W, __mmask8 __U, __m128d __A)
(__mmask8) __U);
}
+extern __inline __m256i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_or_epi64 (__m256i __A, __m256i __B)
+{
+ return (__m256i) ((__v4du)__A | (__v4du)__B);
+}
+
extern __inline __m128i
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_or_epi64 (__m128i __W, __mmask8 __U, __m128i __A, __m128i __B)
(__mmask8) __U);
}
+extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_or_epi64 (__m128i __A, __m128i __B)
+{
+ return (__m128i) ((__v2du)__A | (__v2du)__B);
+}
+
extern __inline __m256i
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm256_mask_xor_epi64 (__m256i __W, __mmask8 __U, __m256i __A,
(__mmask8) __U);
}
+extern __inline __m256i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm256_xor_epi64 (__m256i __A, __m256i __B)
+{
+ return (__m256i) ((__v4du)__A ^ (__v4du)__B);
+}
+
extern __inline __m128i
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_mask_xor_epi64 (__m128i __W, __mmask8 __U, __m128i __A,
(__mmask8) __U);
}
+extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+_mm_xor_epi64 (__m128i __A, __m128i __B)
+{
+ return (__m128i) ((__v2du)__A ^ (__v2du)__B);
+}
+
extern __inline __m256d
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm256_mask_max_pd (__m256d __W, __mmask8 __U, __m256d __A,
--- /dev/null
+/* { dg-do compile { target { ! ia32 } } } */
+/* { dg-options "-O2 -mavx512vl" } */
+/* { dg-final { scan-assembler-times "vpord\[^\n\r\]*ymm16" 1 } } */
+/* { dg-final { scan-assembler-times "vpord\[^\n\r\]*xmm16" 1 } } */
+/* { dg-final { scan-assembler-times "vporq\[^\n\r\]*ymm16" 1 } } */
+/* { dg-final { scan-assembler-times "vporq\[^\n\r\]*xmm16" 1 } } */
+/* { dg-final { scan-assembler-times "vpxord\[^\n\r\]*ymm16" 1 } } */
+/* { dg-final { scan-assembler-times "vpxord\[^\n\r\]*xmm16" 1 } } */
+/* { dg-final { scan-assembler-times "vpxorq\[^\n\r\]*ymm16" 1 } } */
+/* { dg-final { scan-assembler-times "vpxorq\[^\n\r\]*xmm16" 1 } } */
+
+#include <immintrin.h>
+
+__m256i
+foo1 (__m256i x, __m256i y)
+{
+ register __m256i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm256_or_epi32 (x, z);
+}
+
+__m256i
+foo2 (__m256i x, __m256i y)
+{
+ register __m256i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm256_xor_epi32 (x, z);
+}
+
+__m128i
+foo3 (__m128i x, __m128i y)
+{
+ register __m128i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm_or_epi32 (x, z);
+}
+
+__m128i
+foo4 (__m128i x, __m128i y)
+{
+ register __m128i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm_xor_epi32 (x, z);
+}
+
+__m256i
+foo5 (__m256i x, __m256i y)
+{
+ register __m256i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm256_or_epi64 (x, z);
+}
+
+__m256i
+foo6 (__m256i x, __m256i y)
+{
+ register __m256i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm256_xor_epi64 (x, z);
+}
+
+__m128i
+foo7 (__m128i x, __m128i y)
+{
+ register __m128i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm_or_epi64 (x, z);
+}
+
+__m128i
+foo8 (__m128i x, __m128i y)
+{
+ register __m128i z __asm ("xmm16") = y;
+ asm volatile ("" : "+v" (z));
+ return _mm_xor_epi64 (x, z);
+}