[PATCH][AArch64] Custom lower concat_vector patterns with v4i16, v4i32, v8i8, v8i16, v16i8 types.

[Chad Rosier]

Thanks, Ana.  Committed r200491.

> Chad,
> Thanks for fixing these pattern failures.
> LGTM.
> Ana.
>
> -----Original Message-----
> From: Chad Rosier [mailto:mcrosier at codeaurora.org]
> Sent: Thursday, January 30, 2014 7:37 AM
> To: llvm-commits at cs.uiuc.edu
> Cc: t.p.northover at gmail.com; apazos at codeaurora.org; Hao.Liu at arm.com;
> kevinqindev at gmail.com; Jiangning.Liu at arm.com
> Subject: [PATCH][AArch64] Custom lower concat_vector patterns with v4i16,
> v4i32, v8i8, v8i16, v16i8 types.
>
> All,
> The attached patch custom lowers concat_vector patterns with v4i16, v4i32,
> v8i8, v8i16, v16i8 types that have the same operand of type v1* using the
> DUP instruction.  In the case of a constant operand the VDUP, rather than
> VDUPLANE, is used because it avoids an INS instruction.
>
>  Chad
>
> --
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted
> by
> The Linux Foundation
>
>


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