<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=us-ascii">
<meta name="Generator" content="Microsoft Word 14 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
{font-family:Tahoma;
panose-1:2 11 6 4 3 5 4 4 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri","sans-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;
font-family:"Calibri","sans-serif";
color:windowtext;}
span.EmailStyle18
{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="color:#1F497D">I take it back.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="color:#1F497D">The trunk version of LLVM does vectorize the loop. As a result, the microbenchmark also runs quite faster when compiled with clang.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="color:#1F497D"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="color:#1F497D"><o:p> </o:p></span></p>
<div>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in 0in 0in">
<p class="MsoNormal"><b><span style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">From:</span></b><span style="font-size:10.0pt;font-family:"Tahoma","sans-serif""> llvmdev-bounces@cs.uiuc.edu [mailto:llvmdev-bounces@cs.uiuc.edu]
<b>On Behalf Of </b>Murali, Sriram<br>
<b>Sent:</b> Monday, April 15, 2013 11:22 AM<br>
<b>To:</b> llvmdev@cs.uiuc.edu<br>
<b>Subject:</b> [LLVMdev] State of Loop Unrolling and Vectorization in LLVM<o:p></o:p></span></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Hi , I have a test case (and a micro benchmark made out of the test case) to check if loop unrolling and loop vectorization is efficiently done on LLVM. Here is the test case (credits: Tyler Nowicki)<o:p></o:p></p>
<p class="MsoNormal">{code}<o:p></o:p></p>
<p class="MsoNormal">extern float * array;<o:p></o:p></p>
<p class="MsoNormal">extern int array_size;<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">float g()<o:p></o:p></p>
<p class="MsoNormal">{<o:p></o:p></p>
<p class="MsoNormal"> int i;<o:p></o:p></p>
<p class="MsoNormal"> float total = 0;<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"> for(i = 0; i < array_size; i++)<o:p></o:p></p>
<p class="MsoNormal"> {<o:p></o:p></p>
<p class="MsoNormal"> total += array[i];<o:p></o:p></p>
<p class="MsoNormal"> }<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"> return total;<o:p></o:p></p>
<p class="MsoNormal">}<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">{code}<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">When compiled with the options -m32 -mfpmath=sse -ffast-math -funroll-loops -O3 -march=atom for gcc, and clang, I am not able to see the loop being unrolled / vectorized. The microbenchmark which runs the function g() over a billion times
shows quite some performance difference on gcc against clang<o:p></o:p></p>
<p class="MsoNormal">Gcc – 8.6 seconds<o:p></o:p></p>
<p class="MsoNormal">Clang – 12.7 seconds<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Evidently, the addition operation can be vectorized to use addps, (clang does addss), and the loop can be unrolled for better performance. Any idea why this is happening ?<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Thanks<o:p></o:p></p>
<p class="MsoNormal">Sriram<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">--<o:p></o:p></p>
<p class="MsoNormal">Sriram Murali<o:p></o:p></p>
<p class="MsoNormal">SSG/DPD/ECDL/DMP<o:p></o:p></p>
<p class="MsoNormal">+1 (519) 772 – 2579<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</body>
</html>