[polly] Reduction detection [V4]

Johannes Doerfert jdoerfert at codeaurora.org
Tue Jun 17 09:13:56 PDT 2014


Yes, does that mean commit?

--

Johannes Doerfert
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by
The Linux Foundation


-----Original Message-----
From: Tobias Grosser [mailto:tobias at grosser.es] 
Sent: Monday, June 16, 2014 10:53 PM
To: Johannes Doerfert
Cc: llvm-commits at cs.uiuc.edu; 'Sebastian Pop'
Subject: Re: [polly] Reduction detection [V4]

On 17/06/2014 01:29, Johannes Doerfert wrote:
> Hey Tobias,
>
> I didn't consider that the same pointer check would make so much 
> description/comments invalid.
>
> Would you also agree to the patch if:
>    1) I check __not__ for __identical__ pointers but identical base 
> values (like the base address in the memory access).

That works.

>    2) Do not touch the comments [which are more valuable to me than 
> this check].

I just attached a patch that reworks the comments such that they make sense,
but that we still keep the check (that simplifies the dependence review). I
also updated the one test case, where we now do not detect a reduction. Does
this one work for you?

Cheers,
Tobias





More information about the llvm-commits mailing list