<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=iso-8859-1">
<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;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0cm;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri","sans-serif";
mso-fareast-language:EN-US;}
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-compose;
font-family:"Calibri","sans-serif";
color:windowtext;}
.MsoChpDefault
{mso-style-type:export-only;
font-family:"Calibri","sans-serif";
mso-fareast-language:EN-US;}
@page WordSection1
{size:612.0pt 792.0pt;
margin:72.0pt 72.0pt 72.0pt 72.0pt;}
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="DE" link="blue" vlink="purple">
<div class="WordSection1">
<p class="MsoNormal"><span lang="EN-US">Do the fast math flags (esp. nnan) currently have any effect on fcmp instruction optimization?<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><br>
This small module:<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">target triple = "x86_64-pc-windows-msvc"<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">define float @test(float %a, float %b) {<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">entry:<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %0 = fcmp nnan olt float 0.000000e+00, %a<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %1 = fcmp nnan olt float %a, %b<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %2 = and i1 %0, %1<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> br i1 %2, label %entry.t0, label %entry.f0<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">entry.t0: ; preds = %entry<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> ret float %a<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">entry.f0: ; preds = %entry<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %3 = fcmp nnan oge float 0.000000e+00, %a<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> br i1 %3, label %entry.t1, label %entry.f1<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">entry.t1: ; preds = %entry.f0<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> ret float 0.000000e+00<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">entry.f1: ; preds = %entry.f0<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> ret float %b<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">}<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">Currently optimizes to this:<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">target datalayout = "e-m:w-i64:64-f80:128-n8:16:32:64-S128"<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">target triple = "x86_64-pc-windows-msvc"<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">; Function Attrs: norecurse nounwind readnone<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">define float @test(float %a, float %b) #0 {<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">entry:<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %0 = fcmp nnan ogt float %a, 0.000000e+00<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %1 = fcmp nnan olt float %a, %b<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %2 = and i1 %0, %1<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %3 = fcmp nnan ole float %a, 0.000000e+00<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %.b = select i1 %3, float 0.000000e+00, float %b<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> %merge = select i1 %2, float %a, float %.b<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"> ret float %merge<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">}<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">attributes #0 = { norecurse nounwind readnone }<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span lang="EN-US">If nans cannot occur, fcmp nnan ole float %a, 0 (or ule, which should be identical without nans) should just be not fcmp nnan ogt float %a, 0, which was already computed, or am I missing something?<o:p></o:p></span></p>
<p class="MsoNormal"><span lang="EN-US"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">--<o:p></o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">Johannes S. Mueller-Roemer, MSc<o:p></o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">Wiss. Mitarbeiter - Interactive Engineering Technologies (IET)<o:p></o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE"><o:p> </o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">Fraunhofer-Institut für Graphische Datenverarbeitung IGD<o:p></o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">Fraunhoferstr. 5 | 64283 Darmstadt | Germany<o:p></o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">Tel +49 6151 155-606 | Fax +49 6151 155-139<o:p></o:p></span></p>
<p class="MsoNormal"><span style="mso-fareast-language:DE">johannes.mueller-roemer@igd.fraunhofer.de | www.igd.fraunhofer.de<o:p></o:p></span></p>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</body>
</html>