[PATCH][AArch64] implement aarch64 neon instruction class AdvSIMD (3 diff)

Jiangning Liu liujiangning1 at gmail.com
Tue Aug 27 00:21:23 PDT 2013


Tim,

Since you already have a patch for 32-bit ARM to remove those builtins,
could you please upstream your code first? And then I can merge your code.

I think this should be the correct order to make this happen. Otherwise, I
would have to make the change for 32-bit and 64-bit simultaneously, because
we are sharing arm_neon.h/arm_neon.td for 32-bit and 64-bit.

Thanks,
-Jiangning


2013/8/26 Tim Northover <t.p.northover at gmail.com>

> > + vqdmlal and vqmlsl are almost certainly find as intrinsics for now.
>
> I wonder if I should finish my thinking before making posts?
> Aaaanyway, I think we can make an vqdmlal out of a vqdmull and a vqadd
> intrinsic. I'll get a patch for that ready.
>
> Tim.
>



-- 
Thanks,
-Jiangning
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130827/f1a9751d/attachment.html>


More information about the llvm-commits mailing list