Load address calculation when using AVX2 gather instructions

by

Gather instructions do not have any alignment requirements. So it would be too restrictive not to allow byte addressing.

Other reason is consistency. With SIB addressing we obviously have byte address:

MOV eax, [rcx + rdx * 2]

Since VPGATHERDD is just a vectorized variant of this MOV instruction, we should not expect anything different with VSIB addressing:

VPGATHERDD ymm0, [rcx + ymm2 * 2], ymm3

As for real life use for byte addressing, we could have a 24-bit color image where each pixel is 3-byte aligned. We could load 8 pixels with single VPGATHERDD instruction but only if “scale” field in VSIB is “1” and VPGATHERDD uses byte addressing.

More Related Contents:

  1. How to efficiently convert an 8-bit bitmap to array of 0/1 integers with x86 SIMD [duplicate]
  2. Find the first instance of a character using simd
  3. Fastest way to compute absolute value using SSE
  4. Fastest Implementation of Exponential Function Using AVX
  5. Header files for x86 SIMD intrinsics
  6. Sum reduction of unsigned bytes without overflow, using SSE2 on Intel
  7. Fastest way to unpack 32 bits to a 32 byte SIMD vector
  8. Convention for displaying vector registers
  9. Fastest way to do horizontal vector sum with AVX instructions [duplicate]
  10. SSE multiplication of 4 32-bit integers
  11. AVX2 what is the most efficient way to pack left based on a mask?
  12. What are the best instruction sequences to generate vector constants on the fly?
  13. is there an inverse instruction to the movemask instruction in intel avx2?
  14. How to implement atoi using SIMD?
  15. What is the meaning of “non temporal” memory accesses in x86
  16. How to merge a scalar into a vector without the compiler wasting an instruction zeroing upper elements? Design limitation in Intel’s intrinsics?
  17. How to perform the inverse of _mm256_movemask_epi8 (VPMOVMSKB)?
  18. How do I enable SSE for my freestanding bootable code?
  19. Loading 8 chars from memory into an __m256 variable as packed single precision floats
  20. Per-element atomicity of vector load/store and gather/scatter?
  21. Getting started with Intel x86 SSE SIMD instructions
  22. Difference between MOVDQA and MOVAPS x86 instructions?
  23. Can I use the AVX FMA units to do bit-exact 52 bit integer multiplications?
  24. Where is VPERMB in AVX2?
  25. Efficient sse shuffle mask generation for left-packing byte elements
  26. Compare 16 byte strings with SSE
  27. How to convert 32-bit float to 8-bit signed char? (4:1 packing of int32 to int8 __m256i)
  28. inlining failed in call to always_inline ‘_mm_mullo_epi32’: target specific option mismatch
  29. Fastest Implementation of the Natural Exponential Function Using SSE
  30. Bubble sort in x86 (masm32), the sort I wrote doesn’t work

Leave a Comment