<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta name="Generator" content="Microsoft Word 15 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
{font-family:Wingdings;
panose-1:5 0 0 0 0 0 0 0 0 0;}
@font-face
{font-family:"MS Mincho";
panose-1:2 2 6 9 4 2 5 8 3 4;}
@font-face
{font-family:"Cambria Math";
panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:"\@MS Mincho";
panose-1:2 2 6 9 4 2 5 8 3 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:12.0pt;
font-family:"Times New Roman",serif;}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{mso-style-priority:99;
color:purple;
text-decoration:underline;}
span.EmailStyle17
{mso-style-type:personal-reply;
font-family:"Calibri",sans-serif;
color:#1F497D;}
.MsoChpDefault
{mso-style-type:export-only;
font-size:10.0pt;}
@page WordSection1
{size:8.5in 11.0in;
margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang="EN-US" link="blue" vlink="purple">
<div class="WordSection1">
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal">>There needs to be some actual motivating case to make it worth even writing the code for.<span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">This goes back into “priority to implement” question. If there aren’t any customers, priority goes down, by a lot.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal">>So under that paradigm - followed religiously - one would plug in any loop transformation, polyhedral or non-polyhedral etc cost models etc to morph code vectorizable<span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">I won’t comment on other transformations. Powerful vectorizer certainly helps other optimizers make a case,<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">and sometimes require more optimizations to fully appreciate.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><a name="_MailEndCompose">>This might be a good paradigm to follow from the peak performance angle, but not so from the compile-time or code size angle.<br>
>It seems best to pursue a paradigm like this with a peak performance library rather than mainstream llvm.</a><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">This should be evaluated feature-by-feature. I fully understand that LLVM is also used as JIT compiler.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">I don’t think FP induction is adding significantly more compile-time and code-size than integer induction.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal">>So i suggest y'all start from: "Here are the cases we care about making faster, and why we care about making them faster”.<o:p></o:p></p>
<p class="MsoNormal">>+1<o:p></o:p></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">This was our thinking before the paradigm shift.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">The following code vectorizes for TTT being int (might need a bit of extension in SCEV) but not when TTT is float/double (unless FP induction<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">analysis is available). Adding 2-lines of code like this to a 1000-line loop suddenly stops vectorizing the entire loop. These are the things that<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">greatly irritate programmers. Resolving programmer frustration is equally important as performance. In this case, a robust vectorizer should<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">either 1) vectorize FP induction or 2) tell the programmer that FP induction needs to be converted to integer induction. Either way, FP induction<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">analysis is needed. Showing a backward dependence edge on “x” would certainly help, but not as helpful as 1) or 2). ICC Vectorizer customers<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">appreciate improved “loop was not vectorized” messaging as much as functional and performance enhancements of the vectorizer.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">In general, investing in making vectorizer “robust” pays off very well, through performance and/or programmer satisfaction.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">void foo(TTT *a, int N, TTT x, TTT y){<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"> int i;<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"> for (i=0;i<N;i++){<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"> A[i] = x;<br>
x+=y;<br>
}<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">}<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">FYI, I have a customer asking for an extension of OpenMP linear for non-POD types (I won’t bother getting into that discussion in llvm-dev).<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">When vectorizer becomes stronger, more feature requests will come.
</span><span style="font-size:11.0pt;font-family:Wingdings;color:#1F497D">J</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">Thanks,<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">Hideki<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"><o:p> </o:p></span></p>
<div>
<div style="border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0in 0in 0in">
<p class="MsoNormal"><a name="_____replyseparator"></a><b><span style="font-size:11.0pt;font-family:"Calibri",sans-serif">From:</span></b><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"> ghoflehner@apple.com [mailto:ghoflehner@apple.com]
<br>
<b>Sent:</b> Tuesday, May 17, 2016 7:07 PM<br>
<b>To:</b> Daniel Berlin <dberlin@dberlin.org>; Saito, Hideki <hideki.saito@intel.com><br>
<b>Cc:</b> llvm-dev <llvm-dev@lists.llvm.org><br>
<b>Subject:</b> Re: [llvm-dev] Working on FP SCEV Analysis<o:p></o:p></span></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<p class="MsoNormal">On May 17, 2016, at 6:14 PM, Daniel Berlin via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>> wrote:<o:p></o:p></p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal">On Tue, May 17, 2016 at 5:17 PM, Saito, Hideki via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:<o:p></o:p></p>
<blockquote style="border:none;border-left:solid #CCCCCC 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-top:5.0pt;margin-right:0in;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"> </span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif">>What situations are they common in?<o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"> </span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">ICC Vectorizer made a paradigm shift a while ago.</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal" style="text-indent:.5in"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">If there aren’t a clear reason why something can’t be vectorized, we should try our best to vectorize.</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">The rest is a performance modeling (and priority to implement) question, not a capability question.</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">We believe this is a good paradigm to follow in a vectorizer development.</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
</div>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">In some sense, yes, but not at all possible costs.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal">There needs to be some actual motivating case to make it worth even writing the code for.<o:p></o:p></p>
</div>
</div>
</div>
</div>
</div>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<p class="MsoNormal">This paradigm can have far reaching consequences. The vectorizer is the performance cow to milk at the IR level. So under that paradigm - followed religiously - one would plug in any loop transformation, polyhedral or non-polyhedral etc
cost models etc to morph code vectorizable. And when that is not sufficient one would probably start adding large numbers of run-time checks, multi-versioned code etc. This might be a good paradigm to follow from the peak performance angle, but not so from
the compile-time or code size angle. It seems best to pursue a paradigm like this with a peak performance library rather than mainstream llvm.<o:p></o:p></p>
</div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<blockquote style="border:none;border-left:solid #CCCCCC 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-top:5.0pt;margin-right:0in;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">It was a big departure from</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">“vectorize when all things look nice to vectorizer”.</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
</div>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">These are not diametrically opposed.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">I mean, it may be not worth the cost of mainintaing the *compiler code* to do o it.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal">This isn't the same as "when things look nice to the vectorizer", it's more "we're willing to vectorize whatever we can, as long as someone is going to actually use it".<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Nobody has here provided a useful set of cases/applications/etc that suggests it should be done. I'm not saying there are none, i'm saying, literally, nobody has motivated this use case yet :)<o:p></o:p></p>
</div>
</div>
</div>
</div>
</div>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<blockquote style="border:none;border-left:solid #CCCCCC 1.0pt;padding:0in 0in 0in 6.0pt;margin-left:4.8pt;margin-top:5.0pt;margin-right:0in;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D"> </span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri",sans-serif;color:#1F497D">We shouldn’t give up vectorizing simply because programmer wrote a FP induction code.(*)</span><span style="font-size:11.0pt;font-family:"Calibri",sans-serif"><o:p></o:p></span></p>
</div>
</div>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">We shouldn't add code to the compiler just because we can.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">I would similarly be against, for example, vectorizing loops with binary coded decimal induction variables, and adding an entire BCD SCEV infrastructure, without some motivating case *somewhere*.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">So i suggest y'all start from: "Here are the cases we care about making faster, and why we care about making them faster”.<o:p></o:p></p>
</div>
</div>
</div>
</div>
</div>
</blockquote>
<p class="MsoNormal">+1 I think a lot of people would be very interested in non-toy examples that show big performance differences between icc and clang. That would also allow to dig deeper into questions like is it “vectorizer capability, dependence analysis
and/or supporting transformations and/or ??? ” to explain the gap.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">In compilers, building infrastructure first, then finding customers works a lot worse than figuring out what customers want, and then building infrastructure for them :)<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</div>
</div>
</div>
<p class="MsoNormal">_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><o:p></o:p></p>
</div>
</blockquote>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</body>
</html>