HA: [PATCH] D13023: [Bug 21683 relative] - Refactor of SimpleDefinedAtom::sortReferences()

George Rimar via llvm-commits llvm-commits at lists.llvm.org
Thu Sep 24 01:59:51 PDT 2015


>От: Rui Ueyama [ruiu at google.com]
>Отправлено: 24 сентября 2015 г. 1:33
>Кому: reviews+D13023+public+b892af2c6a344227 at reviews.llvm.org
>Копия: George Rimar; Nick Kledzik; Lang Hames; Denis Protivensky; llvm-commits; Adhemerval Zanella
>Тема: Re: [PATCH] D13023: [Bug 21683 relative] - Refactor of SimpleDefinedAtom::sortReferences()
>
>>On Wed, Sep 23, 2015 at 12:37 PM, Lang Hames <lhames at gmail.com<mailto:lhames at gmail.com>> wrote:
>>lhames added a subscriber: lhames.
>>lhames added a comment.
>>
>>I would expect the cost of a linked-list sort to quickly exceed the cost of copying to/from the vector and sorting there. If that's the case this tidy-up may not be worth the performance hit?
>
>That's good point and I'd agree. Vectors could be much faster than sorting a linked list in place because of its better locality.
>http://reviews.llvm.org/D13023

Ok than, I agree that vectors could be faster. Not sure if this place really perfomance critical but now I think it really looks to be not worth to make such changes.






More information about the llvm-commits mailing list