Enable vectorization of call instructions in the loop vectorizer

James Molloy James.Molloy at arm.com
Thu Dec 19 11:45:37 PST 2013


Hi Renato,

Valid question! Originally yes that was the intent, but after having tried implementing this interface I realized of course that we need to query if functions are vectorizable in const context.

Therefore, the has* functions are able to be const, and the get* are not and take a Module& to code generate into.

Sent from my Windows Phone
________________________________
From: Renato Golin<mailto:renato.golin at linaro.org>
Sent: ‎19/‎12/‎2013 19:31
To: James Molloy<mailto:James.Molloy at arm.com>
Cc: Hal Finkel<mailto:hfinkel at anl.gov>; Arnold Schwaighofer<mailto:aschwaighofer at apple.com>; llvm-commits<mailto:llvm-commits at cs.uiuc.edu>
Subject: Re: RFC: Enable vectorization of call instructions in the loop vectorizer

On 19 December 2013 18:12, James Molloy <james.molloy at arm.com<mailto:james.molloy at arm.com>> wrote:
Sorry for the churn – one last version! This one adds predicate functions “has*”, and adds a Module& parameter to the “get*” functions to allow them to create the function on demand if needs be.

Hi James,

I may be wrong, but I'd assume the has* functions would return get* != NULL, so not really value added.

Are you implementing them in a different way?

cheers,
--renato

-- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2557590
ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No: 2548782
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20131219/d449e6b0/attachment.html>


More information about the llvm-commits mailing list