<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:Helvetica;
panose-1:2 11 6 4 2 2 2 2 2 4;}
@font-face
{font-family:Helvetica;
panose-1:2 11 6 4 2 2 2 2 2 4;}
@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.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";}
span.BalloonTextChar
{mso-style-name:"Balloon Text Char";
mso-style-priority:99;
mso-style-link:"Balloon Text";
font-family:"Tahoma","sans-serif";}
span.apple-converted-space
{mso-style-name:apple-converted-space;}
span.EmailStyle20
{mso-style-type:personal;
font-family:"Calibri","sans-serif";
color:#1F497D;}
span.EmailStyle21
{mso-style-type:personal;
font-family:"Calibri","sans-serif";
color:#1F497D;}
span.EmailStyle22
{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 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="EN-GB" link="blue" vlink="purple">
<div class="WordSection1">
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">Hi Quentin,<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">One more reproducer, this time with small (<64bit) values being passed on the stack:<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"> int foo(int x0, int x1, int x2, int x3, int x4, int x5, int x6, int x7,<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> int stack1) {<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> return stack1;<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"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> int main() {<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> int ret = foo(0,1,2,3,4,5,6,7,8);<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> printf("%d\n", ret);<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"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">Global isel thinks that the incoming value of stack1 is stored in bytes [0,4) above SP, but for big-endian targets this should be in bytes [4,8):<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">// /work/llvm/build/bin/clang --target=aarch64-arm-none-eabi -march=armv8-a -c callees.cpp -O0 -Wall -std=c++11 -mllvm -global-isel=true -mllvm -global-isel-abort=0
-mbig-endian -o - -S<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">_Z3fooiiiiiiiii: // @_Z3fooiiiiiiiii<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">// BB#0: // %entry<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> sub sp, sp, #48 // =48<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ldr w8, [sp, #48] // <= Should be [sp, #52]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w0, [sp, #44]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w1, [sp, #40]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w2, [sp, #36]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w3, [sp, #32]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w4, [sp, #28]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w5, [sp, #24]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w6, [sp, #20]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w7, [sp, #16]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str w8, [sp, #12]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ldr w0, [sp, #12]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> add sp, sp, #48 // =48<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ret<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">Oliver<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 style="border:none;border-left:solid blue 1.5pt;padding:0cm 0cm 0cm 4.0pt">
<div>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<p class="MsoNormal"><b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">From:</span></b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif""> Oliver Stannard
<br>
<b>Sent:</b> 17 November 2017 14:57<br>
<b>To:</b> 'qcolombet@apple.com'<br>
<b>Cc:</b> 'llvm-dev@lists.llvm.org'; nd; Kristof Beyls<br>
<b>Subject:</b> RE: [llvm-dev] [GlobalISel][AArch64] Toward flipping the switch for O0: Please give it a try!<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">Hi Quentin,<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">It seems that we also get the calling convention wrong for vector types on big-endian:<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #include <arm_neon.h><o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> int32x2_t load_vector(int32x2_t *p) {<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> return *p;<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"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">Global-isel generates this:<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> // armclang --target=aarch64-arm-none-eabi -march=armv8-a -c callees.cpp -O0 -Wall -std=c++11 -mllvm -global-isel=true -mllvm -global-isel-abort=0 -mbig-endian
-o - -S<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> _Z11load_vectorP11__Int32x2_t: // @_Z11load_vectorP11__Int32x2_t<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> // BB#0: // %entry<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> sub sp, sp, #16 // =16<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str x0, [sp, #8]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ldr x0, [sp, #8]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ld1 { v0.2s }, [x0]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> add sp, sp, #16 // =16<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ret<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">With global-isel off, there is a rev64 instruction between the ld1 and the add, which fixes up the endianness of the vector.<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">Oliver<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 style="border:none;border-left:solid blue 1.5pt;padding:0cm 0cm 0cm 4.0pt">
<div>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<p class="MsoNormal"><b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">From:</span></b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif""> Oliver Stannard
<br>
<b>Sent:</b> 17 November 2017 13:32<br>
<b>To:</b> 'qcolombet@apple.com'<br>
<b>Cc:</b> <a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>; nd; Kristof Beyls<br>
<b>Subject:</b> RE: [llvm-dev] [GlobalISel][AArch64] Toward flipping the switch for O0: Please give it a try!<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">Hi Quentin,<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">At Kristof’s suggestion, I tried running our ABI test suite for a big-endian AArch64 target, and this found an ABI mismatch between global-isel and regular
-O0. Here’s a reproducer for the first one I’ve investigated:<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"> struct foo {<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> float first;<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> float second;<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"> float get_first(foo p) {<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> return p.first;<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"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">This is the code that global-isel currently generates:<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> // /work/llvm/build/bin/clang --target=aarch64--none-eabi -march=armv8-a -c callees.cpp -O0 -mllvm -global-isel=true -mllvm -global-isel-abort=0 -mbig-endian
-o - -S<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"> _Z9get_first3foo: // @_Z9get_first3foo<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> // BB#0: // %entry<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> sub sp, sp, #16 // =16<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> // implicit-def: %X8<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> fmov w9, s0<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> mov w10, w9<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> bfxil x8, x10, #0, #32<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> fmov w9, s1<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> mov w10, w9<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> bfi x8, x10, #32, #32<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> add x10, sp, #8 // =8<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> str x8, [sp, #8]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ldr w9, [x10]<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> fmov s0, w9<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> add sp, sp, #16 // =16<o:p></o:p></span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ret<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">When run on a big-endian target, this incorrectly returns the second member of the struct, instead of the first.<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">Oliver<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 style="border:none;border-left:solid blue 1.5pt;padding:0cm 0cm 0cm 4.0pt">
<div>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<p class="MsoNormal"><b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">From:</span></b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">
<a href="mailto:qcolombet@apple.com">qcolombet@apple.com</a> [<a href="mailto:qcolombet@apple.com">mailto:qcolombet@apple.com</a>]
<br>
<b>Sent:</b> 14 November 2017 23:11<br>
<b>To:</b> Quentin Colombet<br>
<b>Cc:</b> Oliver Stannard; <a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>; Justin Bogner; Ahmed Bougacha; Aditya Nandakumar; nd<br>
<b>Subject:</b> Re: [llvm-dev] [GlobalISel][AArch64] Toward flipping the switch for O0: Please give it a try!<o:p></o:p></span></p>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">To give an update here, we actually are not missing a mapping. The code complains because we are copying around a fp16 into a gpr32 and that shouldn’t be done with a copy (default mapping).<o:p></o:p></p>
<div>
<p class="MsoNormal">I extended the repairing code to issue G_ANYEXT in those cases instead of asserting.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">However, now, I have to teach instruction select about those ANYEXT otherwise we’ll fallback in that case. But that’s a different story.<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
<div>
<p class="MsoNormal">I’ll try to commit today or tomorrow (I have to strengthen the tests).<o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<p class="MsoNormal">On Nov 14, 2017, at 9:29 AM, Quentin Colombet via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a>> wrote:<o:p></o:p></p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<p class="MsoNormal">Thanks Oliver.<o:p></o:p></p>
<div>
<p class="MsoNormal">I’ll have a look. This typically means that we miss a mapping for this type/instruction, which is not surprising given how little code we have we fp16.<o:p></o:p></p>
<div>
<div>
<p class="MsoNormal" style="margin-bottom:12.0pt"><o:p> </o:p></p>
<div>
<p class="MsoNormal">On Nov 14, 2017, at 2:27 AM, Oliver Stannard <<a href="mailto:oliver.stannard@arm.com">oliver.stannard@arm.com</a>> wrote:<o:p></o:p></p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">Hi Quentin,</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">I’ve started running an ABI test suite with global isel on AArch64, and while it hasn’t found any ABI issues it has hit an assertion in clang when using the
__fp16 type. Here’s a reproducer:</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> __fp16 pass_f16(__fp16 p) {</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> return p;</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> }</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> $ /work/llvm/build/bin/clang --target=aarch64-arm-none-eabi -march=armv8-a -c test.c -O0 -mllvm -global-isel -mllvm -global-isel-abort=0</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: /work/llvm/llvm/lib/CodeGen/GlobalISel/RegisterBankInfo.cpp:446: static void llvm::RegisterBankInfo::applyDefaultMapping(const llvm::RegisterBankInfo::OperandsMapper
&): Assertion `OrigTy.getSizeInBits() == NewTy.getSizeInBits() && "Types with difference size cannot be handled by the default " "mapping"' failed.</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #0 0x000000000362a764 PrintStackTraceSignalHandler(void*) (/work/llvm/build/bin/clang-6.0+0x362a764)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #1 0x000000000362aac6 SignalHandler(int) (/work/llvm/build/bin/clang-6.0+0x362aac6)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #2 0x00007f9193b78330 __restore_rt (/lib/x86_64-linux-gnu/libpthread.so.0+0x10330)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #3 0x00007f919276bc37 gsignal /build/eglibc-oGUzwX/eglibc-2.19/signal/../nptl/sysdeps/unix/sysv/linux/raise.c:56:0</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #4 0x00007f919276f028 abort /build/eglibc-oGUzwX/eglibc-2.19/stdlib/abort.c:91:0</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #5 0x00007f9192764bf6 __assert_fail_base /build/eglibc-oGUzwX/eglibc-2.19/assert/assert.c:92:0</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #6 0x00007f9192764ca2 (/lib/x86_64-linux-gnu/libc.so.6+0x2fca2)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #7 0x0000000003d70eb9 (/work/llvm/build/bin/clang-6.0+0x3d70eb9)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #8 0x0000000003d6b00c llvm::RegBankSelect::applyMapping(llvm::MachineInstr&, llvm::RegisterBankInfo::InstructionMapping const&, llvm::SmallVectorImpl<llvm::RegBankSelect::RepairingPlacement>&)
(/work/llvm/build/bin/clang-6.0+0x3d6b00c)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #9 0x0000000003d6b366 llvm::RegBankSelect::assignInstr(llvm::MachineInstr&) (/work/llvm/build/bin/clang-6.0+0x3d6b366)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #10 0x0000000003d6b7f1 llvm::RegBankSelect::runOnMachineFunction(llvm::MachineFunction&) (/work/llvm/build/bin/clang-6.0+0x3d6b7f1)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #11 0x0000000002d934c8 llvm::MachineFunctionPass::runOnFunction(llvm::Function&) (/work/llvm/build/bin/clang-6.0+0x2d934c8)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #12 0x00000000030c998f llvm::FPPassManager::runOnFunction(llvm::Function&) (/work/llvm/build/bin/clang-6.0+0x30c998f)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #13 0x00000000030c9c53 llvm::FPPassManager::runOnModule(llvm::Module&) (/work/llvm/build/bin/clang-6.0+0x30c9c53)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #14 0x00000000030ca136 llvm::legacy::PassManagerImpl::run(llvm::Module&) (/work/llvm/build/bin/clang-6.0+0x30ca136)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #15 0x00000000037c3dcf clang::EmitBackendOutput(clang::DiagnosticsEngine&, clang::HeaderSearchOptions const&, clang::CodeGenOptions const&, clang::TargetOptions
const&, clang::LangOptions const&, llvm::DataLayout const&, llvm::Module*, clang::BackendAction, std::unique_ptr<llvm::raw_pwrite_stream, std::default_delete<llvm::raw_pwrite_stream> >) (/work/llvm/build/bin/clang-6.0+0x37c3dcf)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #16 0x0000000003d421a0 clang::BackendConsumer::HandleTranslationUnit(clang::ASTContext&) (/work/llvm/build/bin/clang-6.0+0x3d421a0)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #17 0x0000000004457376 clang::ParseAST(clang::Sema&, bool, bool) (/work/llvm/build/bin/clang-6.0+0x4457376)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #18 0x0000000003ca6ea0 clang::FrontendAction::Execute() (/work/llvm/build/bin/clang-6.0+0x3ca6ea0)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #19 0x0000000003c1fa31 clang::CompilerInstance::ExecuteAction(clang::FrontendAction&) (/work/llvm/build/bin/clang-6.0+0x3c1fa31)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #20 0x0000000003d3bf4b clang::ExecuteCompilerInvocation(clang::CompilerInstance*) (/work/llvm/build/bin/clang-6.0+0x3d3bf4b)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #21 0x0000000001f85629 cc1_main(llvm::ArrayRef<char const*>, char const*, void*) (/work/llvm/build/bin/clang-6.0+0x1f85629)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #22 0x0000000001f83096 main (/work/llvm/build/bin/clang-6.0+0x1f83096)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #23 0x00007f9192756f45 __libc_start_main /build/eglibc-oGUzwX/eglibc-2.19/csu/libc-start.c:321:0</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> #24 0x0000000001f80029 _start (/work/llvm/build/bin/clang-6.0+0x1f80029)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> Stack dump:</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> 0. Program arguments: /work/llvm/build/bin/clang-6.0 -cc1 -triple aarch64-arm-none-eabi -emit-obj -mrelax-all -disable-free -main-file-name test.c -mrelocation-model
static -mthread-model posix -mdisable-fp-elim -fmath-errno -masm-verbose -mconstructor-aliases -fuse-init-array -target-cpu generic -target-feature +neon -target-abi aapcs -fallow-half-arguments-and-returns -dwarf-column-info -debugger-tuning=gdb -coverage-notes-file
/work/innovation/cctest/test.gcno -resource-dir /work/llvm/build/lib/clang/6.0.0 -O0 -fdebug-compilation-dir /work/innovation/cctest -ferror-limit 19 -fmessage-length 226 -fno-signed-char -fobjc-runtime=gcc -fdiagnostics-show-option -fcolor-diagnostics -mllvm
-global-isel -mllvm -global-isel-abort=0 -o test.o -x c test.c<span class="apple-converted-space"> </span></span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> 1. <eof> parser at end of file</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> 2. Code generation</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> 3. Running pass 'Function Pass Manager' on module 'test.c'.</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> 4. Running pass 'RegBankSelect' on function '@pass_f16'</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: error: unable to execute command: Aborted (core dumped)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: error: clang frontend command failed due to signal (use -v to see invocation)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang version 6.0.0 (<a href="ssh://olista01@ds-gerrit.euhpc.arm.com:29418/armcompiler/clang"><span style="color:purple">ssh://olista01@ds-gerrit.euhpc.arm.com:29418/armcompiler/clang</span></a><span class="apple-converted-space"> </span>aa2b9952ef98a5fe2d47384ef17106855b8bae51)
(<a href="ssh://olista01@ds-gerrit.euhpc.arm.com:29418/armcompiler/llvm"><span style="color:purple">ssh://olista01@ds-gerrit.euhpc.arm.com:29418/armcompiler/llvm</span></a><span class="apple-converted-space"> </span>29f89772107a79b5f2a816d4748ed9c19416c1b6)</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> Target: aarch64-arm-none-eabi</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> Thread model: posix</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> InstalledDir: /work/llvm/build/bin</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: note: diagnostic msg: PLEASE submit a bug report to<span class="apple-converted-space"> </span><a href="http://llvm.org/bugs/"><span style="color:purple">http://llvm.org/bugs/</span></a><span class="apple-converted-space"> </span>and
include the crash backtrace, preprocessed source, and associated run script.</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: note: diagnostic msg:</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ********************</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT:</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> Preprocessed source(s) and associated run script(s) are located at:</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: note: diagnostic msg: /tmp/test-e06964.c</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: note: diagnostic msg: /tmp/test-e06964.sh</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> clang-6.0: note: diagnostic msg:</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> ********************</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D">Oliver</span><o:p></o:p></p>
</div>
<div>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1F497D"> </span><o:p></o:p></p>
</div>
<div style="border:none;border-left:solid blue 1.5pt;padding:0cm 0cm 0cm 4.0pt">
<div>
<div style="border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<div>
<p class="MsoNormal"><b><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">From:</span></b><span class="apple-converted-space"><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif""> </span></span><span lang="EN-US" style="font-size:10.0pt;font-family:"Tahoma","sans-serif"">llvm-dev
[<a href="mailto:llvm-dev-bounces@lists.llvm.org">mailto:llvm-dev-bounces@lists.llvm.org</a>]<span class="apple-converted-space"> </span><b>On Behalf Of<span class="apple-converted-space"> </span></b>Quentin Colombet via llvm-dev<br>
<b>Sent:</b><span class="apple-converted-space"> </span>13 November 2017 18:27<br>
<b>To:</b><span class="apple-converted-space"> </span>Kristof Beyls<br>
<b>Cc:</b><span class="apple-converted-space"> </span>llvm-dev; nd; Ahmed Bougacha; Justin Bogner; Aditya Nandakumar<br>
<b>Subject:</b><span class="apple-converted-space"> </span>Re: [llvm-dev] [GlobalISel][AArch64] Toward flipping the switch for O0: Please give it a try!</span><o:p></o:p></p>
</div>
</div>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<p class="MsoNormal">Hi Kristof,<o:p></o:p></p>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal">On Nov 13, 2017, at 9:10 AM, Kristof Beyls <<a href="mailto:Kristof.Beyls@arm.com"><span style="color:purple">Kristof.Beyls@arm.com</span></a>> wrote:<o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<div>
<div>
<p class="MsoNormal">Hi Quentin,<span class="apple-converted-space"> </span><o:p></o:p></p>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">My only remaining concern is around ABI compatibility.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">The following commit seems to indicate that in the previous round of evaluation, we didn’t find an existing ABI compatibility issue:<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><a href="http://llvm.org/viewvc/llvm-project?view=revision&revision=311388"><span style="color:purple">http://llvm.org/viewvc/llvm-project?view=revision&revision=311388</span></a>.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">I haven’t looked into the details of this issue - so maybe I’m worried over nothing?<o:p></o:p></p>
</div>
</div>
</div>
</div>
</blockquote>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">No, you’re right. The problem with ABI is if you are consistently wrong, then you won’t notice :).<o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal" style="margin-bottom:12.0pt"><o:p> </o:p></p>
</div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">I’m wondering if since then on your side you did any testing around ABI compatibility?<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">E.g. building software where you semi-randomly build some functions through GlobalISel and some functions through DAGISel?<o:p></o:p></p>
</div>
</div>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Justin will look into that. Clang has utility script for that utils/ABITest.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Given we will only be able to check iOS ABI, you may want to follow the same kind of validation on your side.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">I let you sync up with Justin for the method.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Cheers,<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">-Quentin<o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal" style="margin-bottom:12.0pt"><o:p> </o:p></p>
</div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Thanks,<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Kristof<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal">On 8 Nov 2017, at 00:42, Quentin Colombet via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org"><span style="color:purple">llvm-dev@lists.llvm.org</span></a>> wrote:<o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<div>
<div>
<div>
<p class="MsoNormal">Hi all,<span class="apple-converted-space"> </span><o:p></o:p></p>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">I’d like to resurrect this thread and ask if people are on board for enabling this by default for AArch64 O0.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">*** What Changed Since June? ***<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- We added a way to describe the legalization actions for non-power-of-2 <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- We gave a tutorial that covers the best practices to target GlobalISel<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- We improved the TableGen backend to reuse existing SDISel patterns<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- We built and ran huge internal software with GISel<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- We evaluated the performance of GISel and are confident things are in a good shape (with<span class="apple-converted-space"> </span><a href="https://reviews.llvm.org/D39034"><span style="color:purple">https://reviews.llvm.org/D39034</span></a>)
and moving forward would look even better (see the last LLVM Dev talk:<span class="apple-converted-space"> </span><i>GlobalISel: Present, Past, and Future</i><span class="apple-converted-space"> </span>when it is available)<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">*** So What’s he Plan? ***<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- Switch the default instruction selector to GISel for AArch64 at O0<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- Enable the fallback path by default for AArch64 (with warnings enabled when that path is hit)<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- Provide a clang option to turn GISel off<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">What do you think?<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Thanks,<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">-Quentin<o:p></o:p></p>
</div>
</div>
<div>
<div>
<div>
<p class="MsoNormal" style="margin-bottom:12.0pt"><o:p> </o:p></p>
</div>
<div>
<div>
<p class="MsoNormal">On Jun 16, 2017, at 4:43 PM, Quentin Colombet <<a href="mailto:qcolombet@apple.com"><span style="color:purple">qcolombet@apple.com</span></a>> wrote:<o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<div>
<div>
<div>
<p class="MsoNormal">Hi all,<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">We had some internal discussions about flipping the default for O0 and we concluded that we wanted to postpone it.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">*** Why Is That? ***<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">We don’t want to send the wrong message that GlobalISel’s design is set in stone and ready for broader adoption.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">In particular,<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">1. The APIs are still evolving and can still possibly change significantly<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">2. The TableGen backend to reuse the existing SD patterns is still at its early stage<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">3. We want to investigate closely the performance of global-isel (compile-time, runtime, code size, fallbacks)<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">The rationale behind those items is that we want to minimize the pain of moving forward for everybody. We also want the out-of-the-box experience to be pleasant (like all/most of the tablegen patterns just work, we have documentation on
how to target a new backend, etc.) Finally, we want to gain confidence we are going to be able to address the performance issues we have with the current design and if not, derive a plan for that.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">We purposely left out of the conversation what will be the right time and requirements to flip the switch. We want to gather more data first. Your help would be appreciated!<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">*** Short-Term Proposal ***<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">What we would like to do instead short-term is:<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">A. Repurpose or create an option “<span style="font-size:8.5pt;background:white">-aarch64-enable-global-isel-at-O</span>” to enable GISel with fallbacks and warnings enables (i.e., equivalent of -global-isel -global-isel-abort=2)<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">B. Advertise this option in the next open source release to allow compiler enthusiastic to try it and report problems<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">C. Have GISel always built so we can push thing in the right place, MachineVerifier in mind, and stop doing some weird gymnastic<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">What do people think?<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">*** Your Help Is Needed ***<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- Please share your experience in using the GISel APIs and how we can make them better. Moving forward we’ll have those conversations on open source instead of internally/with a narrower audience.<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- Report any performance problem you identify<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">- Propose patches!<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">Cheers,<o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal">-Quentin<o:p></o:p></p>
</div>
</div>
<div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal">On Jun 16, 2017, at 3:06 PM, Quentin Colombet via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org"><span style="color:purple">llvm-dev@lists.llvm.org</span></a>> wrote:<o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
<div>
<div>
<blockquote style="margin-top:5.0pt;margin-bottom:5.0pt">
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""><br>
On Jun 14, 2017, at 7:27 AM, Diana Picus <<a href="mailto:diana.picus@linaro.org"><span style="color:purple">diana.picus@linaro.org</span></a>> wrote:</span><o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
<div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">On 12 June 2017 at 18:54, Diana Picus<span class="apple-converted-space"> </span><<a href="mailto:diana.picus@linaro.org" target="_blank"><span style="color:purple">diana.picus@linaro.org</span></a>><span class="apple-converted-space"> </span>wrote:</span><o:p></o:p></p>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Hi all,</span><o:p></o:p></p>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">I added a buildbot [1] running the test-suite with -O0 -global-isel. It runs into the same 2 timeouts that I reported previously on this thread (paq8p and scimark2). It
would be nice to make it green before flipping the switch.</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">I did some more investigations on a machine similar to the one running the buildbot. For paq8p and scimark2, I get these results for O0:</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">PAQ8p:</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Fast isel: 666.344</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Global isel: 731.384</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">SciMark2-C:</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Fast isel: 463.908</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Global isel: 496.22</span><o:p></o:p></p>
</div>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">The current timeout is 500s (so in this particular case we didn't hit it for scimark2, and it ran successfully to completion). I don't think the difference between FastISel
and GlobalISel is too atrocious, so I would propose increasing the timeout for these 2 benchmarks. I'm not sure if we can do this on a per-bot basis, but I see some precedent for setting custom timeout thresholds for various benchmarks on different architectures
(sometimes with comments that it's done so we can run O0 on that particular benchmark). </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Something along these lines works:</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""><a href="https://reviews.llvm.org/differential/diff/102547/" target="_blank"><span style="color:purple">https://reviews.llvm.org/differential/diff/102547/</span></a></span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">What do you guys think about this approach?</span><o:p></o:p></p>
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Looks reasonable to me.</span><o:p></o:p></p>
</div>
</div>
<div>
<p class="MsoNormal" style="margin-bottom:12.0pt"><o:p> </o:p></p>
</div>
<div>
<div>
<div>
<div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Thanks,</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">Diana</span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif""> </span><o:p></o:p></p>
</div>
</div>
<div>
<div>
<p class="MsoNormal"><span style="font-size:9.0pt;font-family:"Helvetica","sans-serif"">PS: The buildbot is using the Makefiles because that's what our other AArch64 test-suite bots use. Moving all of them to CMake is a transition for another time.</span><o:p></o:p></p>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote>
</div>
</div>
</div>
</div>
<div>
<p class="MsoNormal"> <o:p></o:p></p>
</div>
</div>
</div>
</div>
<div>
<p class="MsoNormal">_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org"><span style="color:purple">llvm-dev@lists.llvm.org</span></a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev"><span style="color:purple">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</span></a><o:p></o:p></p>
</div>
</div>
</blockquote>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</div>
</div>
<p class="MsoNormal">_______________________________________________<br>
LLVM Developers mailing list<br>
<a href="mailto:llvm-dev@lists.llvm.org">llvm-dev@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev</a><o:p></o:p></p>
</div>
</blockquote>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</body>
</html>