<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:x="urn:schemas-microsoft-com:office:excel" xmlns:p="urn:schemas-microsoft-com:office:powerpoint" xmlns:a="urn:schemas-microsoft-com:office:access" xmlns:dt="uuid:C2F41010-65B3-11d1-A29F-00AA00C14882" xmlns:s="uuid:BDC6E3F0-6DA3-11d1-A2A3-00AA00C14882" xmlns:rs="urn:schemas-microsoft-com:rowset" xmlns:z="#RowsetSchema" xmlns:b="urn:schemas-microsoft-com:office:publisher" xmlns:ss="urn:schemas-microsoft-com:office:spreadsheet" xmlns:c="urn:schemas-microsoft-com:office:component:spreadsheet" xmlns:odc="urn:schemas-microsoft-com:office:odc" xmlns:oa="urn:schemas-microsoft-com:office:activation" xmlns:html="http://www.w3.org/TR/REC-html40" xmlns:q="http://schemas.xmlsoap.org/soap/envelope/" xmlns:rtc="http://microsoft.com/officenet/conferencing" xmlns:D="DAV:" xmlns:Repl="http://schemas.microsoft.com/repl/" xmlns:mt="http://schemas.microsoft.com/sharepoint/soap/meetings/" xmlns:x2="http://schemas.microsoft.com/office/excel/2003/xml" xmlns:ppda="http://www.passport.com/NameSpace.xsd" xmlns:ois="http://schemas.microsoft.com/sharepoint/soap/ois/" xmlns:dir="http://schemas.microsoft.com/sharepoint/soap/directory/" xmlns:ds="http://www.w3.org/2000/09/xmldsig#" xmlns:dsp="http://schemas.microsoft.com/sharepoint/dsp" xmlns:udc="http://schemas.microsoft.com/data/udc" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:sub="http://schemas.microsoft.com/sharepoint/soap/2002/1/alerts/" xmlns:ec="http://www.w3.org/2001/04/xmlenc#" xmlns:sp="http://schemas.microsoft.com/sharepoint/" xmlns:sps="http://schemas.microsoft.com/sharepoint/soap/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:udcs="http://schemas.microsoft.com/data/udc/soap" xmlns:udcxf="http://schemas.microsoft.com/data/udc/xmlfile" xmlns:udcp2p="http://schemas.microsoft.com/data/udc/parttopart" xmlns:wf="http://schemas.microsoft.com/sharepoint/soap/workflow/" xmlns:dsss="http://schemas.microsoft.com/office/2006/digsig-setup" xmlns:dssi="http://schemas.microsoft.com/office/2006/digsig" xmlns:mdssi="http://schemas.openxmlformats.org/package/2006/digital-signature" xmlns:mver="http://schemas.openxmlformats.org/markup-compatibility/2006" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns:mrels="http://schemas.openxmlformats.org/package/2006/relationships" xmlns:spwp="http://microsoft.com/sharepoint/webpartpages" xmlns:ex12t="http://schemas.microsoft.com/exchange/services/2006/types" xmlns:ex12m="http://schemas.microsoft.com/exchange/services/2006/messages" xmlns:pptsl="http://schemas.microsoft.com/sharepoint/soap/SlideLibrary/" xmlns:spsl="http://microsoft.com/webservices/SharePointPortalServer/PublishedLinksService" xmlns:Z="urn:schemas-microsoft-com:" xmlns:st="" 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 12 (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: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;}
@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">Aligned. Please review<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"><b><i><span style="color:teal">- Elena</span></i></b><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"><o:p></o:p></span></p>
<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""> Chandler Carruth [mailto:chandlerc@google.com]
<br>
<b>Sent:</b> Thursday, February 02, 2012 11:36<br>
<b>To:</b> Demikhovsky, Elena<br>
<b>Cc:</b> llvm-commits@cs.uiuc.edu<br>
<b>Subject:</b> Re: [llvm-commits] [llvm] r149600 - in /llvm/trunk: lib/Target/X86/X86ISelLowering.cpp lib/Target/X86/X86ISelLowering.h lib/Target/X86/X86InstrFragmentsSIMD.td lib/Target/X86/X86InstrSSE.td test/CodeGen/X86/avx-sext.ll<o:p></o:p></span></p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal">On Thu, Feb 2, 2012 at 1:10 AM, Elena Demikhovsky <<a href="mailto:elena.demikhovsky@intel.com">elena.demikhovsky@intel.com</a>> wrote:<o:p></o:p></p>
<p class="MsoNormal">Author: delena<br>
Date: Thu Feb 2 03:10:43 2012<br>
New Revision: 149600<br>
<br>
URL: <a href="http://llvm.org/viewvc/llvm-project?rev=149600&view=rev" target="_blank">
http://llvm.org/viewvc/llvm-project?rev=149600&view=rev</a><br>
Log:<br>
Optimization for SIGN_EXTEND operation on AVX.<br>
Special handling was added for v4i32 -> v4i64 and v8i16 -> v8i32<br>
extensions.<o:p></o:p></p>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Why was this submitted without review? Your original patch was only mailed to the list 19 hours ago.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">This patch doesn't adhere to LLVM coding conventions very closely at all. Please try to make your code look more like the surrounding code. I've commented a few obvious places below.<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-right:0in">
<p class="MsoNormal">+static SDValue PerformSExtCombine(SDNode *N, SelectionDAG &DAG,<br>
+ TargetLowering::DAGCombinerInfo &DCI,<br>
+ const X86Subtarget *Subtarget) {<br>
+ if (!DCI.isBeforeLegalizeOps())<br>
+ return SDValue();<br>
+<br>
+ if (!Subtarget->hasAVX()) return SDValue();<br>
+<br>
+ // Optimize vectors in AVX mode<br>
+ // Sign extend v8i16 to v8i32 and<br>
+ // v4i32 to v4i64<br>
+ //<br>
+ // Divide input vector into two parts<br>
+ // for v4i32 the shuffle mask will be { 0, 1, -1, -1} {2, 3, -1, -1}<br>
+ // use vpmovsx instruction to extend v4i32 -> v2i64; v8i16 -> v4i32<br>
+ // concat the vectors to original VT<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Why is this comment block indented more than the rest of the code?<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-right:0in">
<p class="MsoNormal">+<br>
+ EVT VT = N->getValueType(0);<br>
+ SDValue Op = N->getOperand(0);<br>
+ EVT OpVT = Op.getValueType();<br>
+ DebugLoc dl = N->getDebugLoc();<br>
+<br>
+ if (((VT == MVT::v4i64) && (OpVT == MVT::v4i32)) ||<br>
+ ((VT == MVT::v8i32) && (OpVT == MVT::v8i16))) {<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Please don't wrap comparison operators ('==', '>', '<', etc...) in extraneous parentheses. This defeats many warnings Clang and GCC both implement which use redundant parentheses to force explicit syntax for these two constructs:<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">if ((x = y)) { ... } // Here we assign, and use extra parens to indicate it was intended.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal">if (x == y) { ... } // Here we compare, and w/o the parens will get a warning if we typo '==' as '='.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Also, please line up with the open '('s at each level when breaking a line.<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-right:0in">
<p class="MsoNormal">+<br>
+ unsigned NumElems = OpVT.getVectorNumElements();<br>
+ SmallVector<int,8> ShufMask1(NumElems, -1);<br>
+ for (unsigned i=0; i< NumElems/2; i++) ShufMask1[i] = i;<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Please use the spacing and whitespace conventional in LLVM for your loops. spaces on both sides of '=' and '<' are ubiquitous.<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-right:0in">
<p class="MsoNormal">+<br>
+ SDValue OpLo = DAG.getVectorShuffle(OpVT, dl, Op, DAG.getUNDEF(OpVT),<br>
+ ShufMask1.data());<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Please line up with the '('.<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-right:0in">
<p class="MsoNormal">+<br>
+ SmallVector<int,8> ShufMask2(NumElems, -1);<br>
+ for (unsigned i=0; i< NumElems/2; i++) ShufMask2[i] = i+NumElems/2;<br>
+<br>
+ SDValue OpHi = DAG.getVectorShuffle(OpVT, dl, Op, DAG.getUNDEF(OpVT),<br>
+ ShufMask2.data());<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Please line up with the '('.<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-right:0in">
<p class="MsoNormal">+<br>
+ EVT HalfVT = EVT::getVectorVT(*DAG.getContext(), VT.getScalarType(),<br>
+ VT.getVectorNumElements()/2);<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">Please line up with the '('.<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-right:0in">
<p class="MsoNormal">+<br>
+ OpLo = DAG.getNode(X86ISD::VSEXT_MOVL, dl, HalfVT, OpLo);<br>
+ OpHi = DAG.getNode(X86ISD::VSEXT_MOVL, dl, HalfVT, OpHi);<br>
+<br>
+ return DAG.getNode(ISD::CONCAT_VECTORS, dl, VT, OpLo, OpHi);<br>
+ }<br>
+ return SDValue();<br>
+}<br>
+<br>
static SDValue PerformZExtCombine(SDNode *N, SelectionDAG &DAG,<br>
const X86Subtarget *Subtarget) {<br>
// (i32 zext (and (i8 x86isd::setcc_carry), 1)) -><br>
@@ -14886,6 +14936,7 @@<br>
case X86ISD::BT: return PerformBTCombine(N, DAG, DCI);<br>
case X86ISD::VZEXT_MOVL: return PerformVZEXT_MOVLCombine(N, DAG);<br>
case ISD::ZERO_EXTEND: return PerformZExtCombine(N, DAG, Subtarget);<br>
+ case ISD::SIGN_EXTEND: return PerformSExtCombine(N, DAG, DCI, Subtarget);<br>
case ISD::TRUNCATE: return PerformTruncateCombine(N, DAG, DCI);<br>
case X86ISD::SETCC: return PerformSETCCCombine(N, DAG);<br>
case X86ISD::SHUFP: // Handle all target specific shuffles<br>
<br>
Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.h<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.h?rev=149600&r1=149599&r2=149600&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.h?rev=149600&r1=149599&r2=149600&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/lib/Target/X86/X86ISelLowering.h (original)<br>
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.h Thu Feb 2 03:10:43 2012<br>
@@ -219,6 +219,9 @@<br>
// VZEXT_MOVL - Vector move low and zero extend.<br>
VZEXT_MOVL,<br>
<br>
+ // VZEXT_MOVL - Vector move low and sign extend.<o:p></o:p></p>
</blockquote>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">This comment is wrong.<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-right:0in">
<p class="MsoNormal">+ VSEXT_MOVL,<br>
+<br>
// VSHL, VSRL - 128-bit vector logical left / right shift<br>
VSHLDQ, VSRLDQ,<br>
<br>
<br>
Modified: llvm/trunk/lib/Target/X86/X86InstrFragmentsSIMD.td<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrFragmentsSIMD.td?rev=149600&r1=149599&r2=149600&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrFragmentsSIMD.td?rev=149600&r1=149599&r2=149600&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/lib/Target/X86/X86InstrFragmentsSIMD.td (original)<br>
+++ llvm/trunk/lib/Target/X86/X86InstrFragmentsSIMD.td Thu Feb 2 03:10:43 2012<br>
@@ -71,6 +71,9 @@<br>
SDTCisVT<2, v4f32>, SDTCisPtrTy<3>]>>;<br>
def X86vzmovl : SDNode<"X86ISD::VZEXT_MOVL",<br>
SDTypeProfile<1, 1, [SDTCisSameAs<0,1>]>>;<br>
+def X86vsmovl : SDNode<"X86ISD::VSEXT_MOVL",<br>
+ SDTypeProfile<1, 1, [SDTCisVec<0>, SDTCisInt<1>, SDTCisInt<0>]>>;<br>
+<br>
def X86vzload : SDNode<"X86ISD::VZEXT_LOAD", SDTLoad,<br>
[SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>;<br>
def X86vshldq : SDNode<"X86ISD::VSHLDQ", SDTIntShiftOp>;<br>
<br>
Modified: llvm/trunk/lib/Target/X86/X86InstrSSE.td<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrSSE.td?rev=149600&r1=149599&r2=149600&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrSSE.td?rev=149600&r1=149599&r2=149600&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/lib/Target/X86/X86InstrSSE.td (original)<br>
+++ llvm/trunk/lib/Target/X86/X86InstrSSE.td Thu Feb 2 03:10:43 2012<br>
@@ -5478,6 +5478,16 @@<br>
(PMOVZXDQrm addr:$src)>;<br>
}<br>
<br>
+let Predicates = [HasAVX] in {<br>
+def : Pat<(v2i64 (X86vsmovl (v4i32 VR128:$src))), (VPMOVSXDQrr VR128:$src)>;<br>
+def : Pat<(v4i32 (X86vsmovl (v8i16 VR128:$src))), (VPMOVSXWDrr VR128:$src)>;<br>
+}<br>
+<br>
+let Predicates = [HasSSE41] in {<br>
+def : Pat<(v2i64 (X86vsmovl (v4i32 VR128:$src))), (PMOVSXDQrr VR128:$src)>;<br>
+def : Pat<(v4i32 (X86vsmovl (v8i16 VR128:$src))), (PMOVSXWDrr VR128:$src)>;<br>
+}<br>
+<br>
<br>
multiclass SS41I_binop_rm_int4<bits<8> opc, string OpcodeStr, Intrinsic IntId> {<br>
def rr : SS48I<opc, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src),<br>
<br>
Added: llvm/trunk/test/CodeGen/X86/avx-sext.ll<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx-sext.ll?rev=149600&view=auto" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx-sext.ll?rev=149600&view=auto</a><br>
==============================================================================<br>
--- llvm/trunk/test/CodeGen/X86/avx-sext.ll (added)<br>
+++ llvm/trunk/test/CodeGen/X86/avx-sext.ll Thu Feb 2 03:10:43 2012<br>
@@ -0,0 +1,17 @@<br>
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7-avx -mattr=+avx | FileCheck %s<br>
+<br>
+define <8 x i32> @sext_8i16_to_8i32(<8 x i16> %A) nounwind uwtable readnone ssp {<br>
+;CHECK: sext_8i16_to_8i32<br>
+;CHECK: vpmovsxwd<br>
+<br>
+ %B = sext <8 x i16> %A to <8 x i32><br>
+ ret <8 x i32>%B<br>
+}<br>
+<br>
+define <4 x i64> @sext_4i32_to_4i64(<4 x i32> %A) nounwind uwtable readnone ssp {<br>
+;CHECK: sext_4i32_to_4i64<br>
+;CHECK: vpmovsxdq<br>
+<br>
+ %B = sext <4 x i32> %A to <4 x i64><br>
+ ret <4 x i64>%B<br>
+}<br>
<br>
Propchange: llvm/trunk/test/CodeGen/X86/avx-sext.ll<br>
------------------------------------------------------------------------------<br>
svn:executable = *<br>
<br>
<br>
_______________________________________________<br>
llvm-commits mailing list<br>
<a href="mailto:llvm-commits@cs.uiuc.edu">llvm-commits@cs.uiuc.edu</a><br>
<a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a><o:p></o:p></p>
</blockquote>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<font face="monospace">---------------------------------------------------------------------<br>
Intel Israel (74) Limited<br>
<br>
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.</font></body>
</html>