[PATCH] D8943: Calculate vectorization factor using the narrowest type instead of widest type

Cong Hou via llvm-commits llvm-commits at lists.llvm.org
Mon Sep 21 16:54:34 PDT 2015


congh updated this revision to Diff 35321.
congh added a comment.

Update the patch.

To be able to maximize the bandwidth during vectorization, this update provides a new option vectorizer-maximize-bandwidth. When it is turned on, the vectorizer will determine the vectorization factor (VF) using the smallest instead of widest type in the loop. To avoid increasing register pressure too much, estimates of the register usage for different VFs are calculated so that we only choose a VF when its register usage doesn't exceed the number of available registers.


http://reviews.llvm.org/D8943

Files:
  lib/Transforms/Vectorize/LoopVectorize.cpp
  test/Transforms/LoopVectorize/X86/vector_max_bandwidth.ll
  test/Transforms/LoopVectorize/X86/vector_ptr_load_store.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D8943.35321.patch
Type: text/x-patch
Size: 13437 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150921/42fb18a9/attachment.bin>


More information about the llvm-commits mailing list