Commits on May 19, 2024

1. cranelift: Optimize __multi3-style multiplications …

   LLVM's `__multi3` function works by splitting a wide multiplication into
   several narrower ones. This optimization recognizes the algebraic
   identities involved and merges them back into the original wide
   multiply.
   
   This is not yet done but illustrates how part of the optimization can
   work, at least.
   
   Currently, the lower half of the result is optimized into a single
   `imul` instruction, but most of the intermediate values that are
   optimized away there are still used in computing the upper half, so
   elaboration brings them back later.
   
   Fixes bytecodealliance#4077

   

   jameysharp committed May 19, 2024
   Configuration menu

   • View commit details
   • Copy full SHA for 2713ce9
   • Browse repository at this point

   Copy the full SHA

   2713ce9 View commit details

   Browse the repository at this point in the history