<div dir="ltr"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">Tim,</div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small"><br></div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">
Sorry, I'm not sure I understand your point here. See my comments below, and I'm still trying to understand your point.</div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">
<br></div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">Thanks,</div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">-Jiangning</div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">
<br></div><div class="gmail_extra"><div class="gmail_quote">2013/11/22 Tim Northover <span dir="ltr"><<a href="mailto:t.p.northover@gmail.com" target="_blank">t.p.northover@gmail.com</a>></span><br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div class="im">> As I mentioned previously in this email thread, the scenario for this case<br>
> is different from table lookup. We do have instruction directly map to this<br>
> intrinsic directly. The operand conversion between 64-bit and 128-bit should<br>
> be able to be solved by EXTRACT_SUBREG and SUBREG_TO_REG,<br>
<br>
</div>It certainly *can* be, and my complaint isn't about that. I love<br>
EXTRACT and friends since they're often free! </blockquote><div><br></div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">This sounds good, and we are suggesting using EXTRACT_SUBREG and friends, so we are on the same page, right?</div>
</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">My complaint is that<br>
we're effectively introducing an unnecessary set of intrinsics</blockquote><div><br></div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">I don't understand about this. Can you explicitly point out what intrinsics names are really unnecessary?</div>
</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">(OK, so<br>
they have the same base name as ones that are more useful[*], but<br>
they're still extra backend burden) </blockquote><div><br></div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;font-size:small">Do you mean using EXTRACT_SUBREG and friends is a backend burden?</div>
</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">for something that's representable<br>

at the IR level.<br>
<br>
Cheers.<br>
<br>
Tim.<br>
<br>
[*] Though, now that I think of it, even those ought to be expressible<br>
as IR. Something for the future, I suppose.<br>
</blockquote></div><br><br clear="all"><div><br></div>-- <br><div dir="ltr"><font face="courier new, monospace">Thanks,</font><div><font face="courier new, monospace">-Jiangning</font></div></div>
</div></div>