Commits on Nov 22, 2023

1. Tweaks to the SME2 intrinsics …

   This patch makes a few tweaks to the SME2 intrinsics:
   
   *   The type suffix of the SME2 svread* intrinsics needs to be
       explicit (non-optional), since there is no merge input that
       can be used to infer it.
   
   *   The zn argument to the svluti* intrinsics is a collection of
       2-bit and 4-bit quantities, so it doesn't have a natural element
       size or signedness.  It seems better to keep it as svuint8_t for
       all variants.
   
   *   Because of that, there is no argument that implies the return
       type of the svluti* intrinsics, so the type suffix needs to be
       explicit.  Also, since the instruction performs a bag-of-bits
       lookup, it makes sense to have floating-point variants too.
   
   *   The ZA slice forms of svread* and svwrite* are likewise
       bag-of-bits moves, so we can provide alternatives for all
       element types.
   
   *   arm_neon.h shift-by-immediate instructions use an _n suffix
       to indicate that the shift amount is scalar.  arm_sve.h
       carried this across to the full/non-overloaded forms of SVE
       immediate shifts.  It seems worth doing the same here for
       consistency, and to protect against vector-vector forms
       being added in future.

   

   rsandifo-arm committed Nov 22, 2023
   Configuration menu

   • View commit details
   • Copy full SHA for 250b019
   • Browse repository at this point

   Copy the full SHA

   250b019 View commit details

   Browse the repository at this point in the history