<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 15 (filtered medium)">
<style><!--
/* Font Definitions */
@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;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
font-size:11.0pt;
font-family:"Calibri",sans-serif;}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
p.MsoListParagraph, li.MsoListParagraph, div.MsoListParagraph
{mso-style-priority:34;
margin-top:0in;
margin-right:0in;
margin-bottom:0in;
margin-left:.5in;
font-size:11.0pt;
font-family:"Calibri",sans-serif;}
span.EmailStyle18
{mso-style-type:personal-reply;
font-family:"Calibri",sans-serif;
color:windowtext;}
.MsoChpDefault
{mso-style-type:export-only;
font-family:"Calibri",sans-serif;}
@page WordSection1
{size:8.5in 11.0in;
margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
{page:WordSection1;}
/* List Definitions */
@list l0
{mso-list-id:2023319129;
mso-list-type:hybrid;
mso-list-template-ids:1870271422 67698703 67698713 67698715 67698703 67698713 67698715 67698703 67698713 67698715;}
@list l0:level1
{mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-.25in;}
@list l0:level2
{mso-level-number-format:alpha-lower;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-.25in;}
@list l0:level3
{mso-level-number-format:roman-lower;
mso-level-tab-stop:none;
mso-level-number-position:right;
text-indent:-9.0pt;}
@list l0:level4
{mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-.25in;}
@list l0:level5
{mso-level-number-format:alpha-lower;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-.25in;}
@list l0:level6
{mso-level-number-format:roman-lower;
mso-level-tab-stop:none;
mso-level-number-position:right;
text-indent:-9.0pt;}
@list l0:level7
{mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-.25in;}
@list l0:level8
{mso-level-number-format:alpha-lower;
mso-level-tab-stop:none;
mso-level-number-position:left;
text-indent:-.25in;}
@list l0:level9
{mso-level-number-format:roman-lower;
mso-level-tab-stop:none;
mso-level-number-position:right;
text-indent:-9.0pt;}
ol
{margin-bottom:0in;}
ul
{margin-bottom:0in;}
--></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" style="word-wrap:break-word">
<div class="WordSection1">
<p class="MsoNormal">Yes, I believe it’s from the function simplification pipeline. It happens after inlining *within* an SCC, but since it’s added to the same CFSCCPassManager as the inliner pass, it runs before the inliner pass runs on the *next* SCC, which
effectively revisits the first SCC by transitive inlines.<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">To be concrete:<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Suppose F calls A, A calls B, B calls A.<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Then A and B together belong to SCC 1.<o:p></o:p></p>
<p class="MsoNormal">And F is alone in SCC 2.<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">The order of events is:<o:p></o:p></p>
<ol style="margin-top:0in" start="1" type="1">
<li class="MsoListParagraph" style="margin-left:0in;mso-list:l0 level1 lfo1">CGSCCPassManager visits SCC 1<o:p></o:p></li><ol style="margin-top:0in" start="1" type="a">
<li class="MsoListParagraph" style="margin-left:0in;mso-list:l0 level2 lfo1">InlinerPass visits SCC 1<o:p></o:p></li></ol>
</ol>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>i.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>Call from A to B is considered for inlining, B’s size is B_large, inline not performed<o:p></o:p></p>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>ii.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>Call from B to A is considered for inlining, A’s size is A_large, inline not performed<o:p></o:p></p>
<ol style="margin-top:0in" start="1" type="1">
<ol style="margin-top:0in" start="2" type="a">
<li class="MsoListParagraph" style="margin-left:0in;mso-list:l0 level2 lfo1">FunctionSimplificationPipeline runs inside its CGSCCToFunctionPassAdaptor inside the CGSCCPassManager<o:p></o:p></li></ol>
</ol>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>i.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>FunctionSimplificationPipeline runs on A, now its size is A_small<o:p></o:p></p>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>ii.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>FunctionSimplificationPipeline runs on B, now its size is B_small<o:p></o:p></p>
<ol style="margin-top:0in" start="2" type="1">
<li class="MsoListParagraph" style="margin-left:0in;mso-list:l0 level1 lfo1">CGSCCPassManager visits SCC 2<o:p></o:p></li><ol style="margin-top:0in" start="1" type="a">
<li class="MsoListParagraph" style="margin-left:0in;mso-list:l0 level2 lfo1">InlinerPass visits SCC 2<o:p></o:p></li></ol>
</ol>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>i.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>Call from F to A is considered, A’s size is A_small, inlining is performed. F picks up a call to B in the inlined copy of A.<o:p></o:p></p>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>ii.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>Call form F to B (added by inline just performed) is considered, B’s size is B_small, inlining is performed, F picks up a call to A in the inlined copy of B <-- this is the
step where we are effectively revisiting the call from A to B<o:p></o:p></p>
<p class="MsoListParagraph" style="margin-left:1.5in;text-indent:-1.5in;mso-text-indent-alt:-9.0pt;mso-list:l0 level3 lfo1">
<![if !supportLists]><span style="mso-list:Ignore"><span style="font:7.0pt "Times New Roman"">
</span>iii.<span style="font:7.0pt "Times New Roman""> </span></span><![endif]>Call from F to A (added by inline just performed) is considered, A’s size is A_small, inlining looks profitable but InlineHistory check blocks it<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">In the small example with just one SCC having just two nodes, the existing InlineHistory check is sufficient to curtail the problem. But in our suites that started failing, instead of just one such SCC, there were ~30 (i.e. A and B each
call each other but also call C and D; C and D call each other but also call E and F; E and F each call each other but also call G and H; and so on), which gave the inliner so many apparently-profitable call chains to inline that it ran until OOM.<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">In the reduced repro attached to the bug, each SCC has six nodes, so the code growth is unreasonable with just two or three such SCCs.<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">-Joseph<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<div style="border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0in 0in 0in">
<p class="MsoNormal"><b>From:</b> Arthur Eubanks <aeubanks@google.com> <br>
<b>Sent:</b> Tuesday, June 8, 2021 6:40 PM<br>
<b>To:</b> Joseph Tremoulet <jotrem@microsoft.com><br>
<b>Cc:</b> llvm-dev <llvm-dev@lists.llvm.org><br>
<b>Subject:</b> [EXTERNAL] Re: [llvm-dev] exponential code explosion during inlining<o:p></o:p></p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<p class="MsoNormal">I may be misunderstanding, but when you say "cleanup", do you mean the function simplification pipeline (since the change that caused this was a change to the function simplification pipeline)? But that happens after inlining, and we won't
revisit the SCC, unless there are some weird SCC splitting/merging things going on.<o:p></o:p></p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<p class="MsoNormal">On Tue, Jun 8, 2021 at 8:43 AM Joseph Tremoulet via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>> wrote:<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">
<div>
<div>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Hi,<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I filed a bug (PR50485 [1]) a couple weeks ago for some pathological behavior we’ve hit in the inliner, but there hasn’t been any reply on the bug so I figured I’d broaden the audience
and ask about the issue here.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">The problem is partially a phase-ordering issue – we run cleanup optimizations after inlining an SCC that reduce the size of functions in the SCC, so a given call foo -> bar that
looks unprofitable due to bar’s size while processing the SCC containing foo and bar may suddenly look profitable due to bar’s reduced size when later considering a copy of that same callsite that gets created by inlining foo into some function in a later
SCC.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">This interacts badly with the approach that the inliner relies on local heuristics to eventually converge (rather than limiting itself with some global budget). I’ll copy the comment
explaining that approach here:<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// We use a single common worklist for calls across the entire SCC. We<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// process these in-order and append new calls introduced during inlining to<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// the end.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
//<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// Note that this particular order of processing is actually critical to<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// avoid very bad behaviors. Consider *highly connected* call graphs where<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// each function contains a small amount of code and a couple of calls to<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// other functions. Because the LLVM inliner is fundamentally a bottom-up<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// inliner, it can handle gracefully the fact that these all appear to be<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// reasonable inlining candidates as it will flatten things until they become<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// too big to inline, and then move on and flatten another batch.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
//<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// However, when processing call edges *within* an SCC we cannot rely on this<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// bottom-up behavior. As a consequence, with heavily connected *SCCs* of<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// functions we can end up incrementally inlining N calls into each of<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// N functions because each incremental inlining decision looks good and we<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// don't have a topological ordering to prevent explosions.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
//<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// To compensate for this, we don't process transitive edges made immediate<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// by inlining until we've done one pass of inlining across the entire SCC.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// Large, highly connected SCCs still lead to some amount of code bloat in<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// this model, but it is uniformly spread across all the functions in the SCC<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// and eventually they all become too large to inline, rather than<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto;margin-left:.5in">
// incrementally maknig a single function grow in a super linear fashion.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">The problem in a nutshell is that “eventually they all become too large to inline” is true while inlining is happening in their SCC, but then the cleanup makes them small again
and so the inliner goes nuts chasing all the now-profitable paths through the highly connected SCC when considering them as transitive inlines into a subsequent SCC.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">I’d love some thoughts on how we might best go about addressing this. I could imagine trying to address it as a phase ordering issue by running cleanup at the start of inlining
an SCC – in the cases where we’ve hit this the cleanup hasn’t actually depended on inlining to expose the opportunities, it just happened to first get cleaned up immediately post inlining. I could also imagine trying to address it by limiting transitive inlines
at callsites created by inlining functions from already-converged SCCs, which we could either do wholesale (if we’re expecting them to be too big to inline at this point, that shouldn’t be a big loss, right?) or just by capping their depth, say, to cut off
exponential explosion.<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">Thanks,<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">-Joseph<o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto"> <o:p></o:p></p>
<p class="MsoNormal" style="mso-margin-top-alt:auto;mso-margin-bottom-alt:auto">[1] -
<a href="https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbugs.llvm.org%2Fshow_bug.cgi%3Fid%3D50485&data=04%7C01%7Cjotrem%40microsoft.com%7C9b5731388bd54a08905a08d92ace56b1%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637587888412026728%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C2000&sdata=ordnapD5P%2BPcaPDqKKIgRfNiPbqQEOK8aYjq2uADpzA%3D&reserved=0" target="_blank">
50485 – Exponential code explosion during inlining (llvm.org)</a><o:p></o:p></p>
</div>
</div>
<p class="MsoNormal">_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a><br>
<a href="https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.llvm.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fllvm-dev&data=04%7C01%7Cjotrem%40microsoft.com%7C9b5731388bd54a08905a08d92ace56b1%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C637587888412026728%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C2000&sdata=5cyNhNOo1Bf3Cf8tgygi%2F318Be7oyFpRSj4%2F3UpTAhw%3D&reserved=0" target="_blank">https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><o:p></o:p></p>
</blockquote>
</div>
</div>
</body>
</html>