问题
So, AVX has a function from immintrin.h, which should allow to store the concatenation of two __m128i values into a single __m256i value. The function is
__m256i _mm256_set_m128i (__m128i hi, __m128i lo)
However, when I use it, like so:
__m256i as[2]; __m128i s[4]; as[0] = _mm256_setr_m128i(s[0], s[1]);
I get a compilation error:
error: incompatible types when assigning to type ‘__m256i’ from type ‘int’
I don't really understand why this happens. Any help is greatly appreciated!
回答1:
Not all compilers seem to have _mm256_setr_m128i, or even _mm256_set_m128i, defined in immintrin.h. So I usually just define macros as needed, bracketed with suitable #ifdefs which test for compiler and version:
#define _mm256_set_m128i(v0, v1) _mm256_insertf128_si256(_mm256_castsi128_si256(v1), (v0), 1)
#define _mm256_setr_m128i(v0, v1) _mm256_set_m128i((v1), (v0))
Intel ICC 11.1 and later has both
_mm256_set_m128iand_mm256_setr_m128i.MSVC 2012 and later has just
_mm256_set_m128i.gcc/clang don't seem to have either, although I haven't checked recent versions to see if this has been fixed yet.
回答2:
We had the same problem and used a macro to work around it.
#ifdef __GNUC__
#if __GNUC__ < 8
#define _mm256_set_m128i(xmm1, xmm2) _mm256_permute2f128_si256(_mm256_castsi128_si256(xmm1), _mm256_castsi128_si256(xmm2), 2)
#define _mm256_set_m128f(xmm1, xmm2) _mm256_permute2f128_ps(_mm256_castps128_ps256(xmm1), _mm256_castps128_ps256(xmm2), 2)
#endif
#endif
来源：https://stackoverflow.com/questions/32630458/setting-m256i-to-the-value-of-two-m128i-values