<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 14 (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:Wingdings;
        panose-1:5 0 0 0 0 0 0 0 0 0;}
@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:0cm;
        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;}
p
        {mso-style-priority:99;
        mso-margin-top-alt:auto;
        margin-right:0cm;
        mso-margin-bottom-alt:auto;
        margin-left:0cm;
        font-size:12.0pt;
        font-family:"Times New Roman","serif";}
p.MsoAcetate, li.MsoAcetate, div.MsoAcetate
        {mso-style-priority:99;
        mso-style-link:"Balloon Text Char";
        margin:0cm;
        margin-bottom:.0001pt;
        font-size:8.0pt;
        font-family:"Tahoma","sans-serif";}
p.emailquote, li.emailquote, div.emailquote
        {mso-style-name:emailquote;
        mso-margin-top-alt:auto;
        margin-right:0cm;
        mso-margin-bottom-alt:auto;
        margin-left:1.0pt;
        border:none;
        padding:0cm;
        font-size:12.0pt;
        font-family:"Times New Roman","serif";}
span.BalloonTextChar
        {mso-style-name:"Balloon Text Char";
        mso-style-priority:99;
        mso-style-link:"Balloon Text";
        font-family:"Tahoma","sans-serif";}
span.EmailStyle21
        {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:612.0pt 792.0pt;
        margin:72.0pt 90.0pt 72.0pt 90.0pt;}
div.WordSection1
        {page:WordSection1;}
/* List Definitions */
@list l0
        {mso-list-id:799957587;
        mso-list-type:hybrid;
        mso-list-template-ids:-1214880770 44585902 67698691 67698693 67698689 67698691 67698693 67698689 67698691 67698693;}
@list l0:level1
        {mso-level-start-at:0;
        mso-level-number-format:bullet;
        mso-level-text:-;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:"Calibri","sans-serif";
        mso-fareast-font-family:"Times New Roman";
        mso-bidi-font-family:"Times New Roman";}
@list l0:level2
        {mso-level-number-format:bullet;
        mso-level-text:o;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:"Courier New";
        mso-bidi-font-family:"Times New Roman";}
@list l0:level3
        {mso-level-number-format:bullet;
        mso-level-text:;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:Wingdings;}
@list l0:level4
        {mso-level-number-format:bullet;
        mso-level-text:;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:Symbol;}
@list l0:level5
        {mso-level-number-format:bullet;
        mso-level-text:o;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:"Courier New";
        mso-bidi-font-family:"Times New Roman";}
@list l0:level6
        {mso-level-number-format:bullet;
        mso-level-text:;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:Wingdings;}
@list l0:level7
        {mso-level-number-format:bullet;
        mso-level-text:;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:Symbol;}
@list l0:level8
        {mso-level-number-format:bullet;
        mso-level-text:o;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:"Courier New";
        mso-bidi-font-family:"Times New Roman";}
@list l0:level9
        {mso-level-number-format:bullet;
        mso-level-text:;
        mso-level-tab-stop:none;
        mso-level-number-position:left;
        text-indent:-18.0pt;
        font-family:Wingdings;}
ol
        {margin-bottom:0cm;}
ul
        {margin-bottom:0cm;}
--></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">I wrote a loop with conditional load and store and measured performance on AVX2, where masking support is very basic, relatively to AVX-512.<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">I got 2x speedup with vpmaskmovd.<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 maskmov instruction is slower than one vector load or store, but much faster than 8 scalar memory operations and 8 branches.<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">Usage of masked instructions on AVX-512 will give much more. There is no latency on target in comparison to the regular vector memop.<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>
<p class="MsoNormal" style="margin-left:36.0pt;text-indent:-18.0pt;mso-list:l0 level1 lfo1">
<![if !supportLists]><span style="font-family:"Calibri","sans-serif";color:#31849B"><span style="mso-list:Ignore">-<span style="font:7.0pt "Times New Roman"">         
</span></span></span><![endif]><span dir="LTR"></span><b><i><span style="color:#31849B"> Elena<o:p></o:p></span></i></b></p>
</div>
<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 #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<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""> Das, Dibyendu [mailto:Dibyendu.Das@amd.com]
<br>
<b>Sent:</b> Friday, October 24, 2014 16:20<br>
<b>To:</b> Demikhovsky, Elena; 'llvmdev@cs.uiuc.edu'<br>
<b>Cc:</b> 'dag@cray.com'<br>
<b>Subject:</b> Re: [LLVMdev] Adding masked vector load and store intrinsics<o:p></o:p></span></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">This looks to be a reasonable proposal. However native instructions that support such masked ld/st may have a high latency ? Also, it would be good to state
 some workloads where this will have a positive impact.<br>
<br>
-dibyendu </span><br>
 <o:p></o:p></p>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<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"">: Demikhovsky, Elena [<a href="mailto:elena.demikhovsky@intel.com">mailto:elena.demikhovsky@intel.com</a>]
<br>
<b>Sent</b>: Friday, October 24, 2014 06:24 AM Central Standard Time<br>
<b>To</b>: <a href="mailto:llvmdev@cs.uiuc.edu">llvmdev@cs.uiuc.edu</a> <<a href="mailto:llvmdev@cs.uiuc.edu">llvmdev@cs.uiuc.edu</a>>
<br>
<b>Cc</b>: <a href="mailto:dag@cray.com">dag@cray.com</a> <<a href="mailto:dag@cray.com">dag@cray.com</a>>
<br>
<b>Subject</b>: [LLVMdev] Adding masked vector load and store intrinsics <br>
</span> <o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">Hi,</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"> </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 would like to add support for masked vector loads and stores by introducing new target-independent intrinsics. The loop vectorizer will then be enhanced
 to optimize loops containing conditional memory accesses by generating these intrinsics for existing targets such as AVX2 and AVX-512. The vectorizer will first ask the target about availability of masked vector loads and stores. The SLP vectorizer can potentially
 be enhanced to use these intrinsics as well.</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"> </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 intrinsics would be legal for all targets; targets that do not support masked vector loads or stores will scalarize them.</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 addressed memory will not be touched for masked-off lanes. In particular, if all lanes are masked off no address will be accessed.</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"> </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">  call void @llvm.masked.store (i32* %addr, <16 x i32> %data, i32 4, <16 x i1> %mask)</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"> </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">  %data = call <8 x i32> @llvm.masked.load (i32* %addr, <8 x i32> %passthru, i32 4, <8 x i1> %mask)</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"> </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">where %passthru is used to fill the elements of %data that are masked-off (if any; can be zeroinitializer or undef).</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"> </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">Comments so far, before we dive into more details?</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"> </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">Thank you.</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"> </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">- Elena and Ayal</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"> </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""> <o:p></o:p></span></p>
</div>
<p>---------------------------------------------------------------------<br>
Intel Israel (74) Limited<o:p></o:p></p>
<p>This e-mail and any attachments may contain confidential material for<br>
the sole use of the intended recipient(s). Any review or distribution<br>
by others is strictly prohibited. If you are not the intended<br>
recipient, please contact the sender and delete all copies.<o:p></o:p></p>
</div>
<p>---------------------------------------------------------------------<br>
Intel Israel (74) Limited</p>

<p>This e-mail and any attachments may contain confidential material for<br>
the sole use of the intended recipient(s). Any review or distribution<br>
by others is strictly prohibited. If you are not the intended<br>
recipient, please contact the sender and delete all copies.</p></body>
</html>