[PATCH] D94779: [Clang] Ensure vector predication pragma is ignored only when vectorization width is 1.

Sjoerd Meijer via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Mon Jan 18 05:06:29 PST 2021


SjoerdMeijer added a comment.

In D94779#2504519 <https://reviews.llvm.org/D94779#2504519>, @malharJ wrote:

> I had a look at the Clang Language Extension <https://clang.llvm.org/docs/LanguageExtensions.html#extensions-for-loop-hint-optimizations> ... and I saw this:
>
>> Specifying a width/count of 1 disables the optimization, and is equivalent to vectorize(disable) or interleave(disable).

Ah yes, thanks. I think we should improve the langref spec also, this should be described in the same paragraph, not floating below some example.

With this cleared up, I agree with Florian, ignoring things isn't great. 
But this brings me back to my question earlier: what is the problem with the current behaviour (i.e. not ignoring/handling it)?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D94779/new/

https://reviews.llvm.org/D94779



More information about the cfe-commits mailing list