How to emulate _mm256_loadu_epi32 with gcc or clang?

by

Just use _mm256_loadu_si256 like a normal person. The only thing the AVX512 intrinsic gives you is a nicer prototype (const void* instead of const __m256i*) so you don’t have to write ugly casts.

@chtz suggests out that you might still want to write a wrapper function yourself to get the void* prototype. But don’t call it _mm256_loadu_epi32; some future GCC version will probably add that for compat with Intel’s docs and break your code.

From another perspective, it’s unfortunate that compilers don’t treat it as an AVX1 intrinsic, but I guess compilers which don’t optimize intrinsics, and which let you use intrinsics from ISA extensions you haven’t enabled, need this kind of clue to know when they can use ymm16-31.

You don’t even want the compiler to emit vmovdqu32 ymm when you’re not masking; vmovdqu ymm is shorter and does exactly the same thing, with no penalty for mixing with EVEX-encoded instructions. The compiler can always use an vmovdqu32 or 64 if it wants to load into ymm16..31, otherwise you want it to use a shorter VEX-coded AVX1 vmovdqu.

I’m pretty sure that GCC treats _mm256_maskz_epi32(0xffu,ptr) exactly the same as _mm256_loadu_si256((const __m256i*)ptr) and makes the same asm regardless of which one you use. It can optimize away the 0xffu mask and simply use an unmasked load, but there’s no need for that extra complication in your source.

But unfortunately GCC9 and earlier will pessimize to vmovdqu32 ymm0, [mem] when AVX512VL is enabled (e.g. -march=skylake-avx512) even when you write _mm256_loadu_si256. This was a missed-optimization, GCC Bug 89346.

It doesn’t matter which 256-bit load intrinsic you use (except for aligned vs. unaligned) as long as there’s no masking.

Related:

More Related Contents:

  1. Is `reinterpret_cast`ing between hardware SIMD vector pointer and the corresponding type an undefined behavior?
  2. When should I use _mm_sfence _mm_lfence and _mm_mfence
  3. Intrinsics for CPUID like informations?
  4. inlining failed in call to always_inline ‘__m256d _mm256_broadcast_sd(const double*)’
  5. SSE, intrinsics, and alignment
  6. RDRAND and RDSEED intrinsics on various compilers?
  7. Fastest method to calculate sum of all packed 32-bit integers using AVX512 or AVX2
  8. Emulating shifts on 32 bytes with AVX
  9. C++ error: ‘_mm_sin_ps’ was not declared in this scope
  10. Get member of __m128 by index?
  11. Visual Studio 2017: _mm_load_ps often compiled to movups
  12. How to implement “_mm_storeu_epi64” without aliasing problems?
  13. Most efficient way to check if all __m128i components are 0 [using
  14. VS: unexpected optimization behavior with _BitScanReverse64 intrinsic
  15. C++11 does not deduce type when std::function or lambda functions are involved
  16. How to make cin take only numbers
  17. Why is a pure virtual function initialized by 0?
  18. check variadic templates parameters for uniqueness
  19. How to recover view space position given view space depth value and ndc xy
  20. Can an enum class be converted to the underlying type?
  21. C++ typedef interpretation of const pointers
  22. Converting std::string to std::vector
  23. How to set breakpoints on future shared libraries with a command flag
  24. cmath vs math.h (And similar c-prefixed vs .h extension headers)
  25. Convert Eigen Matrix to C array
  26. std::vector resize downward
  27. Is there some meaningful statistical data to justify keeping signed integer arithmetic overflow undefined?
  28. pointer as non-type template argument
  29. Workaround for error C2536: cannot specify explicit initializer for arrays in Visual Studio 2013
  30. C++ “error: passing ‘const std::map’ as ‘this’ argument of …”

Leave a Comment