intrinsics

Basic C++ class question: I have simple code currently that looks like something like this: typedef int sType; int array[100]; int test(sType s) { return array[ (int)s ]; } What I want, is to convert "sType" to a class, such that the "return array[ (int)s ]" line does not need to be changed. e.g. (pseudocode) class sType { public: int castInt() { return val; } int val; } int array[100]; int test(sType s) { return array[ (int)s ]; } Thanks for any help. class sType { public: operator int() const { return val; } private: int val; }; class sType { public: operator int() const { return val; } int

I am trying to convert a loop I have into a SSE intrinsics. I seem to have made fairly good progress, and by that I mean It's in the correct direction however I appear to have done some of the translation wrong somewhere as I am not getting the same "correct" answer which results from the non-sse code. My original loop which I unrolled by a factor of 4 looks like this: int unroll_n = (N/4)*4; for (int j = 0; j < unroll_n; j++) { for (int i = 0; i < unroll_n; i+=4) { float rx = x[j] - x[i]; float ry = y[j] - y[i]; float rz = z[j] - z[i]; float r2 = rx*rx + ry*ry + rz*rz + eps; float r2inv = 1

this is sort of a follow up on some previous questions on bit manipulation. I modified the code from this site to enumerate strings with K of N bits set (x is the current int64_t with K bits set, and at the end of this code it is the lexicographically next integer with K bits set): int64_t b, t, c, m, r,z; b = x & -x; t = x + b; c = x^t; // was m = (c >> 2)/b per link z = __builtin_ctz(x); m = c >> 2+z; x = t|m; The modification using __builtin_ctz() works fine as long as the least significant one bit is in the lower DWORD of x, but if is not, it totally breaks. This can be seen with the

What is the best practice for swapping __m128i variables? The background is a compile error under Sun Studio 12.2 , which is a C++03 compiler. __m128i is an opaque type used with MMX and SSE instructions, and its usually and unsigned long long[2] . C++03 does not provide the support for swapping arrays, and std:swap(__m128i a, __m128i b) fails under the compiler. Here are some related questions that don't quite hit the mark. They don't apply because std::vector is not available. How can we swap 2 arrays in constant complexity or O(1)? Is it possible to swap arrays of structs C++03 moving a

This question already has an answer here: what's the difference between _mm256_lddqu_si256 and _mm256_loadu_si256 1 answer I would like to know more about the _mm_lddqu_si128 intrinsic ( lddqu instruction since SSE3) particularly compared with the _mm_loadu_si128 intrinsic (movdqu instruction since SSE2) . I only discovered _mm_lddqu_si128 today. The intel intrinsic guide says this intrinsic may perform better than _mm_loadu_si128 when the data crosses a cache line boundary and a comment says it will perform better under certain circumstances, but never perform worse. So why is it not used

I'm trying to test some of the Intel Intrinsics to see how they work. So, i created a function to do that for me and this is the code: void test_intel_256() { __m256 res,vec1,vec2; __M256_MM_SET_PS(vec1, 7.0, 7.0, 7.0, 7.0, 7.0, 7.0, 7.0, 7.0); __M256_MM_SET_PS(vec1, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0); __M256_MM_ADD_PS(res,vec1,vec2); if (res[0] ==9 && res[1] ==9 && res[2] ==9 && res[3] ==9 && res[4] ==9 && res[5] ==9 && res[6] ==9 && res[7] ==9 ) printf("Addition : OK!\n"); else printf("Addition : FAILED!\n"); } But then i'm getting these errors: error: unknown type name ‘__m256’ error: