<div dir="ltr"><div>Hi Cameron, </div><div><br></div>LGTM, thank you!<div>Probably, you will modify instructions in TD file to support masking patterns, I recommend to do these using AVX512_maskable multiclass.<br><div><br></div><div><br></div></div></div><div class="gmail_extra"><br><div class="gmail_quote">2014-11-10 17:39 GMT+03:00 Cameron McInally <span dir="ltr"><<a href="mailto:cameron.mcinally@nyu.edu" target="_blank">cameron.mcinally@nyu.edu</a>></span>:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hey Robert,<br>
<br>
Please find an updated patch attached.<br>
<br>
Also, note that patterns will need to be added to match the masking<br>
shift operations. I will have to address these in a separate patch.<br>
<span class="HOEnZb"><font color="#888888"><br>
-Cameron<br>
</font></span><div class="HOEnZb"><div class="h5"><br>
On Fri, Nov 7, 2014 at 12:19 PM, Robert Khasanov <<a href="mailto:rob.khasanov@gmail.com">rob.khasanov@gmail.com</a>> wrote:<br>
> Hi Cameron,<br>
><br>
> Your AVX512 builtins don't support merge-masking and zero-masking versions.<br>
> General, merge-masked and zero-masked Intel intrinsics will return one<br>
> builtin with the different mask and passthrough args.<br>
> Please add these arguments to your intrinsics, then you will need to create<br>
> another intrinsic type (VSHIFT_MASK), and handle it using<br>
> getVectorMaskingNode() function.<br>
><br>
><br>
> 2014-11-07 2:24 GMT+03:00 Cameron McInally <<a href="mailto:cameron.mcinally@nyu.edu">cameron.mcinally@nyu.edu</a>>:<br>
>><br>
>> Hey guys,<br>
>><br>
>> Here is a small patch to add 512b shift by immediate intrinsics.<br>
>><br>
>> If everything is cool, I'll send a patch for shift by vector next.<br>
>><br>
>> Thanks,<br>
>> Cam<br>
><br>
><br>
</div></div></blockquote></div><br></div>