blob: 11bec09950f1cbb8d334f6efe9a11e047094417c [file] [log] [blame]
Misha Brukman80731b92003-11-22 00:38:41 +00001<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
Reid Spencere4190fa2004-11-18 18:38:58 +00005 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
Misha Brukman80731b92003-11-22 00:38:41 +00006 <link rel="stylesheet" href="llvm.css" type="text/css">
Chad Rosier06da55e2011-05-27 22:50:46 +00007 <title>LLVM 3.0 Release Notes</title>
Misha Brukman80731b92003-11-22 00:38:41 +00008</head>
9<body>
Chris Lattner3d482502003-10-02 04:57:28 +000010
Chad Rosier06da55e2011-05-27 22:50:46 +000011<h1>LLVM 3.0 Release Notes</h1>
Mikhail Glushenkov024f7cf2008-10-13 02:08:34 +000012
Chris Lattnerc871bac2010-03-17 04:02:39 +000013<img align=right src="http://llvm.org/img/DragonSmall.png"
Gabor Greif27b166352010-04-22 10:21:43 +000014 width="136" height="136" alt="LLVM Dragon Logo">
Chris Lattnerc871bac2010-03-17 04:02:39 +000015
Chris Lattner3d482502003-10-02 04:57:28 +000016<ol>
Misha Brukman80731b92003-11-22 00:38:41 +000017 <li><a href="#intro">Introduction</a></li>
Chris Lattnerf5cd9862008-10-13 18:01:01 +000018 <li><a href="#subproj">Sub-project Status Update</a></li>
Chad Rosier06da55e2011-05-27 22:50:46 +000019 <li><a href="#externalproj">External Projects Using LLVM 3.0</a></li>
20 <li><a href="#whatsnew">What's New in LLVM 3.0?</a></li>
Chris Lattner77a51732004-04-30 22:17:12 +000021 <li><a href="GettingStarted.html">Installation Instructions</a></li>
Dan Gohmanad888912008-10-14 16:23:02 +000022 <li><a href="#knownproblems">Known Problems</a></li>
Misha Brukman80731b92003-11-22 00:38:41 +000023 <li><a href="#additionalinfo">Additional Information</a></li>
Chris Lattner3d482502003-10-02 04:57:28 +000024</ol>
25
Chris Lattner020e1fc2004-05-23 21:07:27 +000026<div class="doc_author">
NAKAMURA Takumica46f5a2011-04-09 02:13:37 +000027 <p>Written by the <a href="http://llvm.org/">LLVM Team</a></p>
Misha Brukman80731b92003-11-22 00:38:41 +000028</div>
Chris Lattner3d482502003-10-02 04:57:28 +000029
Chris Lattner17c170a2011-04-06 06:29:50 +000030<!--
Chad Rosier06da55e2011-05-27 22:50:46 +000031<h1 style="color:red">These are in-progress notes for the upcoming LLVM 3.0
Jeffrey Yasskin0830b972010-01-28 01:14:43 +000032release.<br>
33You may prefer the
Chad Rosier06da55e2011-05-27 22:50:46 +000034<a href="http://llvm.org/releases/2.9/docs/ReleaseNotes.html">LLVM 2.9
Dan Gohman62af9d22010-05-03 23:51:05 +000035Release Notes</a>.</h1>
Chris Lattner17c170a2011-04-06 06:29:50 +000036 -->
Jeffrey Yasskin0830b972010-01-28 01:14:43 +000037
Chris Lattner3d482502003-10-02 04:57:28 +000038<!-- *********************************************************************** -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +000039<h2>
Misha Brukman80731b92003-11-22 00:38:41 +000040 <a name="intro">Introduction</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +000041</h2>
Chris Lattner3d482502003-10-02 04:57:28 +000042<!-- *********************************************************************** -->
43
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +000044<div>
Misha Brukman80731b92003-11-22 00:38:41 +000045
Chris Lattner1e4d5bc2008-10-13 17:57:36 +000046<p>This document contains the release notes for the LLVM Compiler
Bill Wendling16133782011-10-26 18:46:16 +000047 Infrastructure, release 3.0. Here we describe the status of LLVM, including
48 major improvements from the previous release and significant known problems.
49 All LLVM releases may be downloaded from
50 the <a href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
Chris Lattner62495762003-10-02 16:38:05 +000051
Chris Lattnerb5bb5972004-12-07 08:04:13 +000052<p>For more information about LLVM, including information about the latest
Bill Wendling16133782011-10-26 18:46:16 +000053 release, please check out the <a href="http://llvm.org/">main LLVM web
54 site</a>. If you have questions or comments,
55 the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM
56 Developer's Mailing List</a> is a good place to send them.</p>
Chris Lattner3d482502003-10-02 04:57:28 +000057
Bill Wendling16133782011-10-26 18:46:16 +000058<p>Note that if you are reading this file from a Subversion checkout or the main
59 LLVM web page, this document applies to the <i>next</i> release, not the
60 current one. To see the release notes for a specific release, please see the
61 <a href="http://llvm.org/releases/">releases page</a>.</p>
Misha Brukman80731b92003-11-22 00:38:41 +000062
63</div>
Chris Lattneraa61f412009-10-13 17:48:04 +000064
Chris Lattnerce6b0472011-04-05 23:22:33 +000065<!-- Features that need text if they're finished for 3.1:
66 ARM EHABI
Chris Lattnera67df2d2010-04-22 06:28:20 +000067 combiner-aa?
Chris Lattneraa61f412009-10-13 17:48:04 +000068 strong phi elim
Chris Lattneraa61f412009-10-13 17:48:04 +000069 loop dependence analysis
Chris Lattner342f9572010-09-29 05:30:03 +000070 CorrelatedValuePropagation
Chris Lattnerce6b0472011-04-05 23:22:33 +000071 lib/Transforms/IPO/MergeFunctions.cpp => consider for 3.1.
Chris Lattner2b8a52e2008-02-10 07:46:44 +000072 -->
Chris Lattner342f9572010-09-29 05:30:03 +000073
Chris Lattnerf5cd9862008-10-13 18:01:01 +000074<!-- *********************************************************************** -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +000075<h2>
Chris Lattnerf5cd9862008-10-13 18:01:01 +000076 <a name="subproj">Sub-project Status Update</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +000077</h2>
Chris Lattnerf5cd9862008-10-13 18:01:01 +000078<!-- *********************************************************************** -->
Chris Lattner625a3d82008-06-08 21:34:41 +000079
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +000080<div>
Bill Wendling16133782011-10-26 18:46:16 +000081
82<p>The LLVM 3.0 distribution currently consists of code from the core LLVM
83 repository (which roughly includes the LLVM optimizers, code generators and
Chris Lattner6d3eeec2011-11-10 20:15:40 +000084 supporting tools), and the Clang repository. In
Bill Wendling16133782011-10-26 18:46:16 +000085 addition to this code, the LLVM Project includes other sub-projects that are
86 in development. Here we include updates on these subprojects.</p>
Chris Lattnerf5cd9862008-10-13 18:01:01 +000087
Chris Lattnerc75fd522008-06-08 21:58:17 +000088<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +000089<h3>
Chris Lattner44c09cd2008-10-13 18:11:54 +000090<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +000091</h3>
Chris Lattnerc75fd522008-06-08 21:58:17 +000092
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +000093<div>
Chris Lattnerc75fd522008-06-08 21:58:17 +000094
Chris Lattner5de7f6e2010-04-26 17:42:18 +000095<p><a href="http://clang.llvm.org/">Clang</a> is an LLVM front end for the C,
Bill Wendling16133782011-10-26 18:46:16 +000096 C++, and Objective-C languages. Clang aims to provide a better user
97 experience through expressive diagnostics, a high level of conformance to
98 language standards, fast compilation, and low memory use. Like LLVM, Clang
99 provides a modular, library-based architecture that makes it suitable for
100 creating or integrating with other development tools. Clang is considered a
101 production-quality compiler for C, Objective-C, C++ and Objective-C++ on x86
102 (32- and 64-bit), and for darwin/arm targets.</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000103
Chad Rosier06da55e2011-05-27 22:50:46 +0000104<p>In the LLVM 3.0 time-frame, the Clang team has made many improvements:</p>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000105
106<ul>
Bill Wendling16133782011-10-26 18:46:16 +0000107 <li>Greatly improved support for building C++ applications, with greater
108 stability and better diagnostics.</li>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000109
Bill Wendling16133782011-10-26 18:46:16 +0000110 <li><a href="http://clang.llvm.org/cxx_status.html">Improved support</a> for
111 the <a href="http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=50372">C++
112 2011</a> standard, including implementations of non-static data member
113 initializers, alias templates, delegating constructors, the range-based
114 for loop, and implicitly-generated move constructors and move assignment
115 operators, among others.</li>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000116
Bill Wendling16133782011-10-26 18:46:16 +0000117 <li>Implemented support for some features of the upcoming C1x standard,
118 including static assertions and generic selections.</li>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000119
Bill Wendling16133782011-10-26 18:46:16 +0000120 <li>Better detection of include and linking paths for system headers and
121 libraries, especially for Linux distributions.</li>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000122
David Chisnall07618782011-11-26 10:56:17 +0000123 <li>Several improvements to Objective-C support, including:
124
125 <ul>
126 <li><a href="http://clang.llvm.org/docs/AutomaticReferenceCounting.html">
127 Automatic Reference Counting</a> (ARC) and an improved memory model
128 cleanly separating object and C memory.</li>
129
130 <li>A migration tool for moving legacy code to ARC</li>
131
132 <li>Better support for data hiding, allowing instance variables to be
133 declared in implementation contexts or class extensions</li>
134 <li>Weak linking for classes</li>
135 <li>Improved static type checking by inferring the return type of methods
136 such as +alloc and -init.</li>
137 </ul>
138
139 Some new features require either the Mac OS X 10.7 / iOS 5 Objective-C
140 runtime, or version 1.6 or later of the GNUstep Objective-C runtime
141 version.</li>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000142
Bill Wendling16133782011-10-26 18:46:16 +0000143 <li>Implemented a number of optimizations in <tt>libclang</tt>, the Clang C
144 interface, to improve the performance of code completion and the mapping
145 from source locations to abstract syntax tree nodes.</li>
Douglas Gregor936c75b2011-10-15 00:48:01 +0000146</ul>
147
Chris Lattnerfd97b882011-04-05 07:19:28 +0000148
Duncan Sandsce5d9ae2011-04-06 08:07:40 +0000149<p>If Clang rejects your code but another compiler accepts it, please take a
Bill Wendling16133782011-10-26 18:46:16 +0000150 look at the <a href="http://clang.llvm.org/compatibility.html">language
151 compatibility</a> guide to make sure this is not intentional or a known
152 issue.</p>
Bill Wendlingef362462008-10-27 09:27:33 +0000153
Chris Lattner44c09cd2008-10-13 18:11:54 +0000154</div>
155
156<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000157<h3>
Duncan Sands1cd78982011-04-04 11:09:08 +0000158<a name="dragonegg">DragonEgg: GCC front-ends, LLVM back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000159</h3>
Chris Lattneraa61f412009-10-13 17:48:04 +0000160
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000161<div>
Bill Wendling16133782011-10-26 18:46:16 +0000162<p><a href="http://dragonegg.llvm.org/">DragonEgg</a> is a
163 <a href="http://gcc.gnu.org/wiki/plugins">gcc plugin</a> that replaces GCC's
Duncan Sandsa294c542011-11-10 18:44:29 +0000164 optimizers and code generators with LLVM's. It works with gcc-4.5 or gcc-4.6,
165 targets the x86-32 and x86-64 processor families, and has been successfully
166 used on the Darwin, FreeBSD, KFreeBSD, Linux and OpenBSD platforms. It fully
167 supports Ada, C, C++ and Fortran. It has partial support for Go, Java, Obj-C
168 and Obj-C++.</p>
Duncan Sands92452b92010-04-02 09:23:15 +0000169
Bill Wendling16133782011-10-26 18:46:16 +0000170<p>The 3.0 release has the following notable changes:</p>
171
Duncan Sandsa294c542011-11-10 18:44:29 +0000172 <li>GCC version 4.6 is now fully supported.</li>
173
174 <li>Patching and building GCC is no longer required: the plugin should work
175 with your system GCC (version 4.5 or 4.6; on Debian/Ubuntu systems the
176 gcc-4.5-plugin-dev or gcc-4.6-plugin-dev package is also needed).</li>
177
178 <li>The <tt>-fplugin-arg-dragonegg-enable-gcc-optzns</tt> option, which runs
179 GCC's optimizers as well as LLVM's, now works much better. This is the
180 option to use if you want ultimate performance! It not yet completely
181 stable: it may cause the plugin to crash.</li>
182
183 <li>The type and constant conversion logic has been almost entirely rewritten,
184 fixing a multitude of obscure bugs.</li>
185
Duncan Sands7f9a0dc2010-09-30 17:37:34 +0000186<ul>
Chad Rosier06da55e2011-05-27 22:50:46 +0000187<!--
188<li></li>
189-->
Duncan Sands7f9a0dc2010-09-30 17:37:34 +0000190</ul>
Chris Lattneraa61f412009-10-13 17:48:04 +0000191
192</div>
193
Chris Lattner120804a2010-10-03 23:49:06 +0000194<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000195<h3>
Chris Lattner120804a2010-10-03 23:49:06 +0000196<a name="compiler-rt">compiler-rt: Compiler Runtime Library</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000197</h3>
Chris Lattner120804a2010-10-03 23:49:06 +0000198
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000199<div>
Bill Wendling16133782011-10-26 18:46:16 +0000200
201<p>The new LLVM <a href="http://compiler-rt.llvm.org/">compiler-rt project</a>
202 is a simple library that provides an implementation of the low-level
203 target-specific hooks required by code generation and other runtime
204 components. For example, when compiling for a 32-bit target, converting a
205 double to a 64-bit unsigned integer is compiled into a runtime call to the
206 "__fixunsdfdi" function. The compiler-rt library provides highly optimized
207 implementations of this and other low-level routines (some are 3x faster than
208 the equivalent libgcc routines).</p>
Chris Lattner120804a2010-10-03 23:49:06 +0000209
Chad Rosier06da55e2011-05-27 22:50:46 +0000210<p>In the LLVM 3.0 timeframe,</p>
Chris Lattner120804a2010-10-03 23:49:06 +0000211
212</div>
Chris Lattneraa61f412009-10-13 17:48:04 +0000213
214<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000215<h3>
Chris Lattner342f9572010-09-29 05:30:03 +0000216<a name="lldb">LLDB: Low Level Debugger</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000217</h3>
Chris Lattner342f9572010-09-29 05:30:03 +0000218
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000219<div>
Chris Lattner342f9572010-09-29 05:30:03 +0000220
Bill Wendling16133782011-10-26 18:46:16 +0000221<p>LLDB has advanced by leaps and bounds in the 3.0 timeframe. It is
222 dramatically more stable and useful, and includes both a
223 new <a href="http://lldb.llvm.org/tutorial.html">tutorial</a> and
224 a <a href="http://lldb.llvm.org/lldb-gdb.html">side-by-side comparison with
225 GDB</a>.</p>
Chris Lattner342f9572010-09-29 05:30:03 +0000226
227</div>
228
229<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000230<h3>
Chris Lattner342f9572010-09-29 05:30:03 +0000231<a name="libc++">libc++: C++ Standard Library</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000232</h3>
Chris Lattner342f9572010-09-29 05:30:03 +0000233
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000234<div>
Chris Lattner342f9572010-09-29 05:30:03 +0000235
Bill Wendling16133782011-10-26 18:46:16 +0000236<p>Like compiler_rt, libc++ is now <a href="DeveloperPolicy.html#license">dual
237 licensed</a> under the MIT and UIUC license, allowing it to be used more
238 permissively.</p>
Chris Lattner342f9572010-09-29 05:30:03 +0000239
David Chisnall07618782011-11-26 10:56:17 +0000240<p>Libc++ has been ported to FreeBSD and imported into the base system. It is
241 planned to be the default STL implementation for FreeBSD 10.</p>
242
Chris Lattner342f9572010-09-29 05:30:03 +0000243</div>
244
Chris Lattneraa61f412009-10-13 17:48:04 +0000245
Chris Lattnercaefe932011-04-07 03:08:22 +0000246<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000247<h3>
Chris Lattnercaefe932011-04-07 03:08:22 +0000248<a name="LLBrowse">LLBrowse: IR Browser</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000249</h3>
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000250
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000251<div>
Bill Wendling16133782011-10-26 18:46:16 +0000252
253<p><a href="http://llvm.org/svn/llvm-project/llbrowse/trunk/doc/LLBrowse.html">
254 LLBrowse</a> is an interactive viewer for LLVM modules. It can load any LLVM
255 module and displays its contents as an expandable tree view, facilitating an
256 easy way to inspect types, functions, global variables, or metadata nodes. It
257 is fully cross-platform, being based on the popular wxWidgets GUI
258 toolkit.</p>
259
Chris Lattnercaefe932011-04-07 03:08:22 +0000260</div>
261
262<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000263<h3>
Chris Lattnercaefe932011-04-07 03:08:22 +0000264<a name="vmkit">VMKit</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000265</h3>
Chris Lattnercaefe932011-04-07 03:08:22 +0000266
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000267<div>
Bill Wendling16133782011-10-26 18:46:16 +0000268
Nicolas Geoffray5e515792011-11-10 23:37:56 +0000269 <p>The <a href="http://vmkit.llvm.org/">VMKit project</a> is an
270 implementation of a Java Virtual Machine (Java VM or JVM) that uses LLVM for
271 static and just-in-time compilation.
272
273 <p>In the LLVM 3.0 time-frame, VMKit has had significant improvements on both
274 runtime and startup performance:</p>
275
276 <ul>
277 <li>Precompilation: by compiling ahead of time a small subset of Java's core
278 library, the startup performance have been highly optimized to the point that
279 running a 'Hello World' program takes less than 30 milliseconds.</li>
280
281 <li>Customization: by customizing virtual methods for individual classes,
282 the VM can statically determine the target of a virtual call, and decide to
283 inline it.</li>
284
285 <li>Inlining: the VM does more inlining than it did before, by allowing more
286 bytecode instructions to be inlined, and thanks to customization. It also
287 inlines GC barriers, and object allocations.</li>
288
289 <li>New exception model: the generated code for a method that does not do
290 any try/catch is not penalized anymore by the eventuality of calling a
291 method that throws an exception. Instead, the method that throws the
292 exception jumps directly to the method that could catch it.</li>
293 </ul>
Bill Wendling16133782011-10-26 18:46:16 +0000294
Chris Lattnercaefe932011-04-07 03:08:22 +0000295</div>
296
297
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000298<!--=========================================================================-->
Chris Lattner9ee0b012011-04-05 18:38:45 +0000299<!--
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000300<h3>
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000301<a name="klee">KLEE: A Symbolic Execution Virtual Machine</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000302</h3>
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000303
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000304<div>
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000305<p>
306<a href="http://klee.llvm.org/">KLEE</a> is a symbolic execution framework for
307programs in LLVM bitcode form. KLEE tries to symbolically evaluate "all" paths
308through the application and records state transitions that lead to fault
309states. This allows it to construct testcases that lead to faults and can even
310be used to verify some algorithms.
311</p>
312
Chris Lattner0d364302011-03-10 07:43:44 +0000313<p>UPDATE!</p>
Chris Lattner9ee0b012011-04-05 18:38:45 +0000314</div>-->
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000315
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000316</div>
Daniel Dunbar8fbd8aa2010-10-04 17:39:47 +0000317
Chris Lattner53e06f92009-02-26 22:33:38 +0000318<!-- *********************************************************************** -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000319<h2>
Chad Rosier06da55e2011-05-27 22:50:46 +0000320 <a name="externalproj">External Open Source Projects Using LLVM 3.0</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000321</h2>
Chris Lattner53e06f92009-02-26 22:33:38 +0000322<!-- *********************************************************************** -->
323
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000324<div>
Chris Lattneraa61f412009-10-13 17:48:04 +0000325
326<p>An exciting aspect of LLVM is that it is used as an enabling technology for
327 a lot of other language and tools projects. This section lists some of the
Chad Rosier06da55e2011-05-27 22:50:46 +0000328 projects that have already been updated to work with LLVM 3.0.</p>
Chris Lattner120804a2010-10-03 23:49:06 +0000329
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000330<!--=========================================================================-->
Bill Wendling0ed57472011-10-26 00:17:54 +0000331<h3>AddressSanitizer</h3>
332
333<div>
334
335<p><a href="http://code.google.com/p/address-sanitizer/">AddressSanitizer</a>
336 uses compiler instrumentation and a specialized malloc library to find C/C++
337 bugs such as use-after-free and out-of-bound accesses to heap, stack, and
338 globals. The key feature of the tool is speed: the average slowdown
339 introduced by AddressSanitizer is less than 2x.</p>
340
341</div>
342
343<!--=========================================================================-->
Bill Wendling3bb971b2011-10-26 00:14:36 +0000344<h3>ClamAV</h3>
345
346<div>
Bill Wendling99ba0b62011-10-25 01:01:42 +0000347
Bill Wendling3bb971b2011-10-26 00:14:36 +0000348<p><a href="http://www.clamav.net">Clam AntiVirus</a> is an open source (GPL)
349 anti-virus toolkit for UNIX, designed especially for e-mail scanning on mail
350 gateways.</p>
351
352<p>Since version 0.96 it
353 has <a href="http://vrt-sourcefire.blogspot.com/2010/09/introduction-to-clamavs-low-level.html">bytecode
354 signatures</a> that allow writing detections for complex malware.</p>
355
356<p>It uses LLVM's JIT to speed up the execution of bytecode on X86, X86-64,
357 PPC32/64, falling back to its own interpreter otherwise. The git version was
358 updated to work with LLVM 3.0.</p>
359
360</div>
361
362<!--=========================================================================-->
Tobias Grosser8bee91f2011-11-14 09:09:26 +0000363<h3>clang_complete for VIM</h3>
364
365<div>
366
367<p><a href="https://github.com/Rip-Rip/clang_complete">clang_complete</a> is a
368 VIM plugin, that provides accurate C/C++ autocompletion using the clang front
369 end. The development version of clang complete, can directly use libclang
370 which can maintain a cache to speed up auto completion.</p>
371
372</div>
373
374<!--=========================================================================-->
Bill Wendling5ee37be2011-10-26 18:23:06 +0000375<h3>clReflect</h3>
376
377<div>
378
379<p><a href="https://bitbucket.org/dwilliamson/clreflect">clReflect</a> is a C++
380 parser that uses clang/LLVM to derive a light-weight reflection database
381 suitable for use in game development. It comes with a very simple runtime
382 library for loading and querying the database, requiring no external
383 dependencies (including CRT), and an additional utility library for object
384 management and serialisation.</p>
385
386</div>
387
388<!--=========================================================================-->
Bill Wendlingca8dc4a2011-10-29 01:10:01 +0000389<h3>Cling C++ Interpreter</h3>
390
391<div>
392
393<p><a href="http://cern.ch/cling">Cling</a> is an interactive compiler interface
394 (aka C++ interpreter). It uses LLVM's JIT and clang; it currently supports
395 C++ and C. It has a prompt interface, runs source files, calls into shared
396 libraries, prints the value of expressions, even does runtime lookup of
397 identifiers (dynamic scopes). And it just behaves like one would expect from
398 an interpreter.</p>
399
400</div>
401
402<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000403<h3>Crack Programming Language</h3>
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000404
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000405<div>
Bill Wendlinged4cc442011-11-03 20:10:01 +0000406
407<p><a href="http://code.google.com/p/crack-language/">Crack</a> aims to provide
408 the ease of development of a scripting language with the performance of a
409 compiled language. The language derives concepts from C++, Java and Python,
410 incorporating object-oriented programming, operator overloading and strong
411 typing.</p>
412
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000413</div>
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000414
415<!--=========================================================================-->
Bill Wendlingb33bd112011-11-08 05:22:54 +0000416<h3>Eero</h3>
417
418<div>
419
420<p><a href="http://eerolanguage.org/">Eero</a> is a fully
421 header-and-binary-compatible dialect of Objective-C 2.0, implemented with a
422 patched version of the Clang/LLVM compiler. It features a streamlined syntax,
423 Python-like indentation, and new operators, for improved readability and
424 reduced code clutter. It also has new features such as limited forms of
425 operator overloading and namespaces, and strict (type-and-operator-safe)
426 enumerations. It is inspired by languages such as Smalltalk, Python, and
427 Ruby.</p>
428
429</div>
430
431<!--=========================================================================-->
Chris Lattnere5b37be2011-11-25 20:28:16 +0000432<h3>FAUST Real-Time Audio Signal Processing Language</h3>
433
434<div>
435
436<p><a href="http://faust.grame.fr/">FAUST</a> is a compiled language for
437 real-time audio signal processing. The name FAUST stands for Functional
438 AUdio STream. Its programming model combines two approaches: functional
439 programming and block diagram composition. In addition with the C, C++, Java
440 output formats, the Faust compiler can now generate LLVM bitcode, and works
441 with LLVM 2.7-3.0.
442 </p>
443
444</div>
445
446<!--=========================================================================-->
Bill Wendling3bb93ad2011-10-26 00:09:55 +0000447<h3>Glasgow Haskell Compiler (GHC)</h3>
448
449<div>
450
451<p>GHC is an open source, state-of-the-art programming suite for Haskell, a
452 standard lazy functional programming language. It includes an optimizing
453 static compiler generating good code for a variety of platforms, together
454 with an interactive system for convenient, quick development.</p>
455
456<p>GHC 7.0 and onwards include an LLVM code generator, supporting LLVM 2.8 and
457 later. Since LLVM 2.9, GHC now includes experimental support for the ARM
458 platform with LLVM 3.0.</p>
459
460</div>
461
462<!--=========================================================================-->
Bill Wendling3e3968e2011-10-25 20:35:31 +0000463<h3>gwXscript</h3>
464
465<div>
466
467<p><a href="http://botwars.tk/gwscript/">gwXscript</a> is an object oriented,
Bill Wendlingc902d132011-10-26 04:24:15 +0000468 aspect oriented programming language which can create both executables (ELF,
Bill Wendling3e3968e2011-10-25 20:35:31 +0000469 EXE) and shared libraries (DLL, SO, DYNLIB). The compiler is implemented in
470 its own language and translates scripts into LLVM-IR which can be optimized
471 and translated into native code by the LLVM framework. Source code in
472 gwScript contains definitions that expand the namespaces. So you can build
473 your project and simply 'plug out' features by removing a file. The remaining
474 project does not leave scars since you directly separate concerns by the
475 'template' feature of gwX. It is also possible to add new features to a
476 project by just adding files and without editing the original project. This
477 language is used for example to create games or content management systems
478 that should be extendable.</p>
479
480<p>gwXscript is strongly typed and offers comfort with its native types string,
481 hash and array. You can easily write new libraries in gwXscript or native
482 code. gwXscript is type safe and users should not be able to crash your
483 program or execute malicious code except code that is eating CPU time.</p>
484
485</div>
486
487<!--=========================================================================-->
Bill Wendlinga2ae87b2011-10-26 22:55:18 +0000488<h3>include-what-you-use</h3>
489
490<div>
491
492<p><a href="http://code.google.com/p/include-what-you-use">include-what-you-use</a>
493 is a tool to ensure that a file directly <code>#include</code>s
494 all <code>.h</code> files that provide a symbol that the file uses. It also
495 removes superfluous <code>#include</code>s from source files.</p>
496
497</div>
498
499<!--=========================================================================-->
Bill Wendlingfc1935c2011-11-07 22:05:17 +0000500<h3>ispc: The Intel SPMD Program Compiler</h3>
501
502<div>
503
504<p><a href="http://ispc.github.com">ispc</a> is a compiler for "single program,
505 multiple data" (SPMD) programs. It compiles a C-based SPMD programming
506 language to run on the SIMD units of CPUs; it often delivers 5-6x speedups on
507 a single core of a CPU with an 8-wide SIMD unit compared to serial code,
508 while still providing a clean and easy-to-understand programming model. For
509 an introduction to the language and its performance,
Chris Lattner6d3eeec2011-11-10 20:15:40 +0000510 see <a href="http://ispc.github.com/example.html">the walkthrough</a> of a short
Bill Wendlingfc1935c2011-11-07 22:05:17 +0000511 example program. ispc is licensed under the BSD license.</p>
512
513</div>
Chris Lattnerc3e4fdc2011-11-25 20:36:17 +0000514
515<!--=========================================================================-->
516<h3>The Julia Programming Language</h3>
517
518<div>
519
520<p><a href="http://github.com/JuliaLang/julia">Julia</a> is a high-level,
521 high-performance dynamic language for technical
522 computing. It provides a sophisticated compiler, distributed parallel
523 execution, numerical accuracy, and an extensive mathematical function
524 library. The compiler uses type inference to generate fast code
525 without any type declarations, and uses LLVM's optimization passes and
526 JIT compiler. The language is designed around multiple dispatch,
527 giving programs a large degree of flexibility. It is ready for use on many
528 kinds of problems.</p>
529</div>
Bill Wendlingfc1935c2011-11-07 22:05:17 +0000530
531<!--=========================================================================-->
Bill Wendling4d9c8e52011-10-26 18:20:54 +0000532<h3>LanguageKit and Pragmatic Smalltalk</h3>
533
534<div>
535
536<p><a href="http://etoileos.com/etoile/features/languagekit/">LanguageKit</a> is
537 a framework for implementing dynamic languages sharing an object model with
538 Objective-C. It provides static and JIT compilation using LLVM along with
539 its own interpreter. Pragmatic Smalltalk is a dialect of Smalltalk, built on
540 top of LanguageKit, that interfaces directly with Objective-C, sharing the
541 same object representation and message sending behaviour. These projects are
Benjamin Kramera02af612011-11-25 21:26:00 +0000542 developed as part of the &Eacute;toil&eacute; desktop environment.</p>
Bill Wendling4d9c8e52011-10-26 18:20:54 +0000543
544</div>
545
546<!--=========================================================================-->
Bill Wendlingeda350d2011-11-01 04:08:23 +0000547<h3>LuaAV</h3>
548
549<div>
550
551<p><a href="http://lua-av.mat.ucsb.edu/blog/">LuaAV</a> is a real-time
552 audiovisual scripting environment based around the Lua language and a
553 collection of libraries for sound, graphics, and other media protocols. LuaAV
554 uses LLVM and Clang to JIT compile efficient user-defined audio synthesis
555 routines specified in a declarative syntax.</p>
556
557</div>
558
559<!--=========================================================================-->
Bill Wendlingfd9eb5732011-10-26 00:16:17 +0000560<h3>Mono</h3>
561
562<div>
563
564<p>An open source, cross-platform implementation of C# and the CLR that is
565 binary compatible with Microsoft.NET. Has an optional, dynamically-loaded
566 LLVM code generation backend in Mini, the JIT compiler.</p>
567
568<p>Note that we use a Git mirror of LLVM with some patches. See:
569 https://github.com/mono/llvm</p>
570
571</div>
572
573<!--=========================================================================-->
Tobias Grossercfa35952011-11-14 09:09:23 +0000574<h3>Polly</h3>
575
576<div>
577
578<p><a href="http://polly.grosser.es">Polly</a> is an advanced data-locality
579 optimizer and automatic parallelizer. It uses an advanced, mathematical
580 model to calculate detailed data dependency information which it uses to
581 optimize the loop structure of a program. Polly can speed up sequential code
582 by improving memory locality and consequently the cache use. Furthermore,
583 Polly is able to expose different kind of parallelism which it exploits by
584 introducing (basic) OpenMP and SIMD code. A mid-term goal of Polly is to
585 automatically create optimized GPU code.</p>
586
587</div>
588
589<!--=========================================================================-->
Bill Wendlinge8caad52011-10-25 20:37:45 +0000590<h3>Portable OpenCL (pocl)</h3>
591
592<div>
593
594<p>Portable OpenCL is an open source implementation of the OpenCL standard which
595 can be easily adapted for new targets. One of the goals of the project is
596 improving performance portability of OpenCL programs, avoiding the need for
597 target-dependent manual optimizations. A "native" target is included, which
598 allows running OpenCL kernels on the host (CPU).</p>
599
600</div>
601
602<!--=========================================================================-->
Bill Wendling58209292011-10-25 20:39:06 +0000603<h3>Pure</h3>
604
605<div>
606<p><a href="http://pure-lang.googlecode.com/">Pure</a> is an
607 algebraic/functional programming language based on term rewriting. Programs
608 are collections of equations which are used to evaluate expressions in a
609 symbolic fashion. The interpreter uses LLVM as a backend to JIT-compile Pure
610 programs to fast native code. Pure offers dynamic typing, eager and lazy
611 evaluation, lexical closures, a hygienic macro system (also based on term
612 rewriting), built-in list and matrix support (including list and matrix
613 comprehensions) and an easy-to-use interface to C and other programming
614 languages (including the ability to load LLVM bitcode modules, and inline C,
615 C++, Fortran and Faust code in Pure programs if the corresponding LLVM-enabled
616 compilers are installed).</p>
617
618<p>Pure version 0.48 has been tested and is known to work with LLVM 3.0
619 (and continues to work with older LLVM releases &gt;= 2.5).</p>
620
621</div>
622
623<!--=========================================================================-->
Bill Wendling97b889d2011-10-26 00:12:04 +0000624<h3>Renderscript</h3>
625
626<div>
627
628<p><a href="http://developer.android.com/guide/topics/renderscript/index.html">Renderscript</a>
629 is Android's advanced 3D graphics rendering and compute API. It provides a
630 portable C99-based language with extensions to facilitate common use cases
631 for enhancing graphics and thread level parallelism. The Renderscript
632 compiler frontend is based on Clang/LLVM. It emits a portable bitcode format
633 for the actual compiled script code, as well as reflects a Java interface for
634 developers to control the execution of the compiled bitcode. Executable
635 machine code is then generated from this bitcode by an LLVM backend on the
636 device. Renderscript is thus able to provide a mechanism by which Android
637 developers can improve performance of their applications while retaining
638 portability.</p>
639
640</div>
641
642<!--=========================================================================-->
Bill Wendlingbf72f2c2011-10-25 20:40:26 +0000643<h3>SAFECode</h3>
644
645<div>
646
647<p><a href="http://safecode.cs.illinois.edu">SAFECode</a> is a memory safe C/C++
648 compiler built using LLVM. It takes standard, unannotated C/C++ code,
649 analyzes the code to ensure that memory accesses and array indexing
650 operations are safe, and instruments the code with run-time checks when
651 safety cannot be proven statically. SAFECode can be used as a debugging aid
652 (like Valgrind) to find and repair memory safety bugs. It can also be used
653 to protect code from security attacks at run-time.</p>
654
655</div>
656
657<!--=========================================================================-->
Bill Wendling8fae82a2011-10-26 07:38:19 +0000658<h3>The Stupid D Compiler (SDC)</h3>
659
660<div>
661
662<p><a href="https://github.com/bhelyer/SDC">The Stupid D Compiler</a> is a
663 project seeking to write a self-hosting compiler for the D programming
664 language without using the frontend of the reference compiler (DMD).</p>
665
666</div>
667
668<!--=========================================================================-->
Bill Wendling53aa7a82011-10-25 20:24:32 +0000669<h3>TTA-based Co-design Environment (TCE)</h3>
670
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000671<div>
Bill Wendling53aa7a82011-10-25 20:24:32 +0000672
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000673<p>TCE is a toolset for designing application-specific processors (ASP) based on
Bill Wendling53aa7a82011-10-25 20:24:32 +0000674 the Transport triggered architecture (TTA). The toolset provides a complete
675 co-design flow from C/C++ programs down to synthesizable VHDL and parallel
676 program binaries. Processor customization points include the register files,
677 function units, supported operations, and the interconnection network.</p>
Bill Wendling16133782011-10-26 18:46:16 +0000678
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000679<p>TCE uses Clang and LLVM for C/C++ language support, target independent
Bill Wendling53aa7a82011-10-25 20:24:32 +0000680 optimizations and also for parts of code generation. It generates new
Bill Wendling16133782011-10-26 18:46:16 +0000681 LLVM-based code generators "on the fly" for the designed TTA processors and
682 loads them in to the compiler backend as runtime libraries to avoid
Bill Wendling53aa7a82011-10-25 20:24:32 +0000683 per-target recompilation of larger parts of the compiler chain.</p>
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000684
Bill Wendling16133782011-10-26 18:46:16 +0000685</div>
Chris Lattnerae6a89a2011-04-06 01:13:49 +0000686
687<!--=========================================================================-->
Bill Wendlingae9aa352011-10-25 20:27:37 +0000688<h3>Tart Programming Language</h3>
689
690<div>
691
692<p><a href="http://code.google.com/p/tart/">Tart</a> is a general-purpose,
693 strongly typed programming language designed for application
694 developers. Strongly inspired by Python and C#, Tart focuses on practical
695 solutions for the professional software developer, while avoiding the clutter
696 and boilerplate of legacy languages like Java and C++. Although Tart is still
697 in development, the current implementation supports many features expected of
698 a modern programming language, such as garbage collection, powerful
699 bidirectional type inference, a greatly simplified syntax for template
700 metaprogramming, closures and function literals, reflection, operator
701 overloading, explicit mutability and immutability, and much more. Tart is
702 flexible enough to accommodate a broad range of programming styles and
703 philosophies, while maintaining a strong commitment to simplicity, minimalism
704 and elegance in design.</p>
705
706</div>
707
708<!--=========================================================================-->
Bill Wendlingfe845d52011-10-26 09:25:01 +0000709<h3>ThreadSanitizer</h3>
710
711<div>
712
713<p><a href="http://code.google.com/p/data-race-test/">ThreadSanitizer</a> is a
714 data race detector for (mostly) C and C++ code, available for Linux, Mac OS
715 and Windows. On different systems, we use binary instrumentation frameworks
Bill Wendlingedc20872011-10-29 01:11:15 +0000716 (Valgrind and Pin) as frontends that generate the program events for the race
717 detection algorithm. On Linux, there's an option of using LLVM-based
718 compile-time instrumentation.</p>
Bill Wendlingfe845d52011-10-26 09:25:01 +0000719
720</div>
721
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000722</div>
723
Chris Lattnerc75fd522008-06-08 21:58:17 +0000724<!-- *********************************************************************** -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000725<h2>
Chad Rosier06da55e2011-05-27 22:50:46 +0000726 <a name="whatsnew">What's New in LLVM 3.0?</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000727</h2>
Chris Lattnerc75fd522008-06-08 21:58:17 +0000728<!-- *********************************************************************** -->
729
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000730<div>
Chris Lattnerb7bc2aa2008-06-08 22:59:35 +0000731
Chris Lattneraa61f412009-10-13 17:48:04 +0000732<p>This release includes a huge number of bug fixes, performance tweaks and
Bill Wendling16133782011-10-26 18:46:16 +0000733 minor improvements. Some of the major improvements and new features are
734 listed in this section.</p>
Chris Lattneraa61f412009-10-13 17:48:04 +0000735
Chris Lattnera67df2d2010-04-22 06:28:20 +0000736<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000737<h3>
Chris Lattnercdc44ed2008-02-06 18:00:06 +0000738<a name="majorfeatures">Major New Features</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000739</h3>
Chris Lattnercdc44ed2008-02-06 18:00:06 +0000740
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000741<div>
Chris Lattnercdc44ed2008-02-06 18:00:06 +0000742
Chris Lattner3dcdc292011-11-27 08:32:32 +0000743<p><b>llvm-gcc is gone</b>. LLVM's configure script doesn't depend on llvm-gcc anymore, clean layering.</p>
Chris Lattner97beb512007-05-14 06:56:09 +0000744
Chris Lattner6d3eeec2011-11-10 20:15:40 +0000745<p>LLVM 3.0 includes several major new capabilities:</p>
746
Chris Lattner92f21832011-11-15 22:13:27 +0000747<!-- Near dead:
748 Analysis/RegionInfo.h + Dom Frontiers
749 SparseBitVector: used in LiveVar.
Chris Lattner3dcdc292011-11-27 08:32:32 +0000750 llvm/lib/Archive - replace with lib object?
Chris Lattner92f21832011-11-15 22:13:27 +0000751 -->
752
753<!--
Chris Lattner74a3e00e2011-11-27 05:47:57 +0000754 Type system rewrite: http://blog.llvm.org/2011/11/llvm-30-type-system-rewrite.html
Chris Lattner92f21832011-11-15 22:13:27 +0000755 Better performance for Neon code in clang due to SRoA improvements.
756 New regalloc on by default. Lin scan going away in 3.1
757 PGO / builtin_expect improvements (summary needed)
758 Big EH rewrite.
759 AVX support, assembler, compiler and disassembler.
760 IndVar improvements: andy
761 PTX backend improvements: Justin
762 llvm-rtdyld & MC JIT: JimG
763 InstAliases now automatically used in the asmprinter where they are shorter.
764 Integrated assembler on by default for arm/thumb?
765 PostOrder Dominator frontiers were removed.
766 Line Profiling / gcov support
767 EH and debug information produced with CFI directives, yielding smaller executables: http://blog.mozilla.com/respindola/2011/05/12/cfi-directives/
768 X86-64 generates smaller and faster code at -O0 (fast isel improvements)
769 Better code generation for Cortex-A9
770 Many APIs take ArrayRef's now.
771 Pass manager extension API.
Chris Lattner74a3e00e2011-11-27 05:47:57 +0000772 ARM inline asm constraints implemented.
773 LangRef.html#fnattrs uwtable attribute for asynch unwind tables.
774 better performance for indirect gotos.
775 llvm.prefetch now takes a 4th argument that specifies whether the prefetch happens to the icache or dcache.
776 New PackedVector, TinyPtrVector class (see Programmer's Manual)
777 New nonlazybind function attribute.
778 ARC language specific optimizer (Transforms/ObjCARC) a decent example of language-specific transformation.
Chris Lattnerbc639292011-11-27 06:56:53 +0000779 LLVM 3.0 removes support for reading LLVM 2.8 and earlier files. Aim to maintain compatibility all the way back to 3.0 "forever".
Chris Lattner614d0392011-11-25 20:33:27 +0000780
Chris Lattner74a3e00e2011-11-27 05:47:57 +0000781 New llvm.expect intrinsic.
782 Table generated MC expansion logic for pseudo instructions that expand to multiple MC instructions through the PseudoInstExpansion class. (JimG)
783 New llvm.fma intrinsic.
784
785 Euro dev meeting and main one too.
Chris Lattnerd5bb9e62011-11-27 07:37:53 +0000786 New atomics instructions, "#i_fence" instruction, cmpxchg, atomicrmw too. What target support (X86/ARM)? Also 'atomic load/store'. See Atomics.html
Chris Lattner74a3e00e2011-11-27 05:47:57 +0000787 X86: inline assembler supports .code32 and .code64.
Chris Lattnerebed15e2011-11-27 06:24:49 +0000788 Exception handling rewrite: new landingpad and resume instruction. Unwind gone.
789 LowerSetJmp pass removed, unused.
Chris Lattner3dcdc292011-11-27 08:32:32 +0000790 llvm-objdump / dwarf parser library / llvm-dwarfdump (d0k)
791 object file parsing stuff and llvm-size (mspencer)
792 llvm-cov (devang)
Chris Lattnerebed15e2011-11-27 06:24:49 +0000793 Old arm disassembler replaced with a new one based on autogenerated encoding information from ARM .td files.
Chris Lattnerd5bb9e62011-11-27 07:37:53 +0000794 Frontend tests removed from llvm/test/Frontend* (was this completed for 3.0?)
795 Segmented stack support (X86 only?) Rafael and Sanjoy Das: docs/SegmentedStacks.html should be in CodeGen.html status table?
796 X86 backend support for NaCl (David Meyer / Nick L)
797 Codegen now supports vector "select" operations on vector comparisons, turning
798 them into various optimized code sequences (e.g. using the SSE4/AVX "blend"
799 instructions).
Chris Lattner3dcdc292011-11-27 08:32:32 +0000800 #line directives in integrated assembler
801 SSE domain fixing code enabled for AVX (Bruno/Jakob). Domain fixing pass is
802 now target independent (ExecutionDepsFix pass). (Jakob)
803 X86 backend synthesizes horizontal add/sub instructions from generic code.
804 returns_twice attribute (rafael)
805 Tablegen has been split into a library, clang tblgen pieces now live in clang.
806 The llvm version is now named llvm-tblgen instead of tblgen.
807 X86: Tons of encoding improvements and new instructions (e.g. Atom, Ivy Bridge,
808 and BMI instructions)
809 added to assembler and disassembler (Craig Topper)
810 data layout string can encode the natural alignment of the target's stack for better optimization (LangRef.html#datalayout)
Chris Lattner92f21832011-11-15 22:13:27 +0000811 -->
812
Chris Lattner97beb512007-05-14 06:56:09 +0000813<ul>
Chris Lattnerc3a2c982011-04-06 00:45:11 +0000814
Chad Rosier06da55e2011-05-27 22:50:46 +0000815<!--
816<li></li>
817-->
Chris Lattner9ee0b012011-04-05 18:38:45 +0000818
Chris Lattner458e79f2008-02-10 08:18:42 +0000819</ul>
Chris Lattnerfd97b882011-04-05 07:19:28 +0000820
Chris Lattner97beb512007-05-14 06:56:09 +0000821</div>
822
Chris Lattner0a1fd102007-09-21 03:54:09 +0000823<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000824<h3>
Chris Lattnerd1094e02009-03-02 02:37:32 +0000825<a name="coreimprovements">LLVM IR and Core Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000826</h3>
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000827
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000828<div>
Bill Wendling16133782011-10-26 18:46:16 +0000829
Chris Lattneraa61f412009-10-13 17:48:04 +0000830<p>LLVM IR has several new features for better support of new targets and that
Bill Wendling16133782011-10-26 18:46:16 +0000831 expose new optimization opportunities:</p>
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000832
Bill Wendlingd535a6e2011-10-26 18:33:01 +0000833<p>One of the biggest changes is that 3.0 has a new exception handling
834 system. The old system used LLVM intrinsics to convey the exception handling
835 information to the code generator. It worked in most cases, but not
836 all. Inlining was especially difficult to get right. Also, the intrinsics
837 could be moved away from the <code>invoke</code> instruction, making it hard
838 to recover that information.</p>
839
840<p>The new EH system makes exception handling a first-class member of the IR. It
841 adds two new instructions:</p>
842
Chris Lattnerb7112222008-06-05 06:25:56 +0000843<ul>
Bill Wendlingd535a6e2011-10-26 18:33:01 +0000844 <li><a href="LangRef.html#i_landingpad"><code>landingpad</code></a> &mdash;
845 this instruction defines a landing pad basic block. It contains all of the
846 information that's needed by the code generator. It's also required to be
847 the first non-PHI instruction in the landing pad. In addition, a landing
848 pad may be jumped to only by the unwind edge of an <code>invoke</code>
849 instruction.</li>
850
851 <li><a href="LangRef.html#i_resume"><code>resume</code></a> &mdash; this
852 instruction causes the current exception to resume traveling up the
853 stack. It replaces the <code>@llvm.eh.resume</code> intrinsic.</li>
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000854</ul>
Mikhail Glushenkov024f7cf2008-10-13 02:08:34 +0000855
Bill Wendlingd535a6e2011-10-26 18:33:01 +0000856<p>Converting from the old EH API to the new EH API is rather simple, because a
857 lot of complexity has been removed. The two intrinsics,
858 <code>@llvm.eh.exception</code> and <code>@llvm.eh.selector</code> have been
Benjamin Kramera02af612011-11-25 21:26:00 +0000859 superseded by the <code>landingpad</code> instruction. Instead of generating
Bill Wendlingd535a6e2011-10-26 18:33:01 +0000860 a call to <code>@llvm.eh.exception</code> and <code>@llvm.eh.selector</code>:
861
862<div class="doc_code">
863<pre>
864Function *ExcIntr = Intrinsic::getDeclaration(TheModule,
865 Intrinsic::eh_exception);
866Function *SlctrIntr = Intrinsic::getDeclaration(TheModule,
867 Intrinsic::eh_selector);
868
869// The exception pointer.
870Value *ExnPtr = Builder.CreateCall(ExcIntr, "exc_ptr");
871
872std::vector&lt;Value*&gt; Args;
873Args.push_back(ExnPtr);
874Args.push_back(Builder.CreateBitCast(Personality,
875 Type::getInt8PtrTy(Context)));
876
877<i>// Add selector clauses to Args.</i>
878
879// The selector call.
880Builder.CreateCall(SlctrIntr, Args, "exc_sel");
881</pre>
882</div>
883
884<p>You should instead generate a <code>landingpad</code> instruction, that
885 returns an exception object and selector value:</p>
886
887<div class="doc_code">
888<pre>
889LandingPadInst *LPadInst =
890 Builder.CreateLandingPad(StructType::get(Int8PtrTy, Int32Ty, NULL),
891 Personality, 0);
892
893Value *LPadExn = Builder.CreateExtractValue(LPadInst, 0);
894Builder.CreateStore(LPadExn, getExceptionSlot());
895
896Value *LPadSel = Builder.CreateExtractValue(LPadInst, 1);
897Builder.CreateStore(LPadSel, getEHSelectorSlot());
898</pre>
899</div>
900
901<p>It's now trivial to add the individual clauses to the <code>landingpad</code>
902 instruction.</p>
903
904<div class="doc_code">
905<pre>
906<i><b>// Adding a catch clause</b></i>
907Constant *TypeInfo = getTypeInfo();
908LPadInst-&gt;addClause(TypeInfo);
909
910<i><b>// Adding a C++ catch-all</b></i>
911LPadInst-&gt;addClause(Constant::getNullValue(Builder.getInt8PtrTy()));
912
913<i><b>// Adding a cleanup</b></i>
914LPadInst-&gt;setCleanup(true);
915
916<i><b>// Adding a filter clause</b></i>
917std::vector&lt;Constant*&gt; TypeInfos;
918Constant *TypeInfo = getFilterTypeInfo();
919TypeInfos.push_back(Builder.CreateBitCast(TypeInfo, Builder.getInt8PtrTy()));
920
921ArrayType *FilterTy = ArrayType::get(Int8PtrTy, TypeInfos.size());
922LPadInst-&gt;addClause(ConstantArray::get(FilterTy, TypeInfos));
923</pre>
924</div>
925
926<p>Converting from using the <code>@llvm.eh.resume</code> intrinsic to
927 the <code>resume</code> instruction is trivial. It takes the exception
928 pointer and exception selector values returned by
929 the <code>landingpad</code> instruction:</p>
930
931<div class="doc_code">
932<pre>
933Type *UnwindDataTy = StructType::get(Builder.getInt8PtrTy(),
934 Builder.getInt32Ty(), NULL);
935Value *UnwindData = UndefValue::get(UnwindDataTy);
936Value *ExcPtr = Builder.CreateLoad(getExceptionObjSlot());
937Value *ExcSel = Builder.CreateLoad(getExceptionSelSlot());
938UnwindData = Builder.CreateInsertValue(UnwindData, ExcPtr, 0, "exc_ptr");
939UnwindData = Builder.CreateInsertValue(UnwindData, ExcSel, 1, "exc_sel");
940Builder.CreateResume(UnwindData);
941</pre>
942</div>
943
Chris Lattnerdd6acc02008-02-10 08:17:19 +0000944</div>
945
946<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000947<h3>
Andrew Trick6c0a11b2011-11-06 17:59:24 +0000948<a name="loopoptimization">Loop Optimization Improvements</a>
949</h3>
950
951<div>
952<p>The induction variable simplification pass in 3.0 only modifies
953 induction variables when profitable. Sign and zero extension
954 elimination, linear function test replacement, loop unrolling, and
955 other simplifications that require induction variable analysis have
956 been generalized so they no longer require loops to be rewritten in a
957 typically suboptimal form prior to optimization. This new design
958 preserves more IR level information, avoids undoing earlier loop
959 optimizations (particularly hand-optimized loops), and no longer
960 strongly depends on the code generator rewriting loops a second time
961 in a now optimal form--an intractable problem.</p>
962
963<p>The original behavior can be restored with -mllvm -enable-iv-rewrite;
964 however, support for this mode will be short lived. As such, bug
965 reports should be filed for any significant performance regressions
966 when moving from -mllvm -enable-iv-rewrite to the 3.0 default mode.</p>
967</div>
968
969<!--=========================================================================-->
970<h3>
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000971<a name="optimizer">Optimizer Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000972</h3>
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000973
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000974<div>
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000975
Chris Lattneraa61f412009-10-13 17:48:04 +0000976<p>In addition to a large array of minor performance tweaks and bug fixes, this
Bill Wendling16133782011-10-26 18:46:16 +0000977 release includes a few major enhancements and additions to the
978 optimizers:</p>
Chris Lattner2b8a52e2008-02-10 07:46:44 +0000979
980<ul>
Benjamin Kramer8c8486d2011-11-26 11:14:54 +0000981<li>Information about <a href="BranchWeightMetadata.html">branch probability</a>
982 and basic block frequency is now available within LLVM, based on a
983 combination of static branch prediction heuristics and
984 <code>__builtin_expect</code> calls. That information is currently used for
985 register spill placement and if-conversion, with additional optimizations
986 planned for future releases. The same framework is intended for eventual
987 use with profile-guided optimization.</li>
Chris Lattner23e16b592011-04-06 05:50:04 +0000988</li>
Chris Lattnerc3a2c982011-04-06 00:45:11 +0000989
Chris Lattnerfcc65a72010-10-04 02:42:39 +0000990</ul>
991
Chris Lattneraa61f412009-10-13 17:48:04 +0000992</div>
Chris Lattnerd1094e02009-03-02 02:37:32 +0000993
Chris Lattner7795ea92008-06-08 02:45:07 +0000994<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000995<h3>
Chris Lattnerf25bc192010-09-30 16:31:33 +0000996<a name="mc">MC Level Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +0000997</h3>
Chris Lattnerf25bc192010-09-30 16:31:33 +0000998
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +0000999<div>
Bill Wendling16133782011-10-26 18:46:16 +00001000
1001<p>The LLVM Machine Code (aka MC) subsystem was created to solve a number of
1002 problems in the realm of assembly, disassembly, object file format handling,
1003 and a number of other related areas that CPU instruction-set level tools work
1004 in.</p>
Chris Lattnerbf1cf672010-10-02 21:59:30 +00001005
Chris Lattnerbf1cf672010-10-02 21:59:30 +00001006<ul>
Jim Grosbach651e2ee2011-11-24 00:49:21 +00001007 <li>The ELF object streamers are much more full featured.</li>
1008 <li>Target dependent relocation handling has been refactored into the Targets.</li>
1009 <li>Early stage MC-JIT infrastructure has been implemented.</li>
Chris Lattnerbf1cf672010-10-02 21:59:30 +00001010</ul>
1011
Jim Grosbach651e2ee2011-11-24 00:49:21 +00001012<p>The MC-JIT is a major new feature for MC, and will eventually grow to replace
1013the current JIT implementation. It emits object files direct to memory and
1014uses a runtime dynamic linker to resolve references and drive lazy compilation.
1015The MC-JIT enables much greater code reuse between the JIT and the static
1016compiler and provides better integration with the platform ABI as a result.</p>
1017
Bill Wendling16133782011-10-26 18:46:16 +00001018<p>For more information, please see
1019 the <a href="http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html">Intro
1020 to the LLVM MC Project Blog Post</a>.</p>
Chris Lattnerf25bc192010-09-30 16:31:33 +00001021
NAKAMURA Takumi8d89b8e2011-04-05 08:24:22 +00001022</div>
Chris Lattnerf25bc192010-09-30 16:31:33 +00001023
Chris Lattnerf25bc192010-09-30 16:31:33 +00001024<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001025<h3>
Chris Lattnerd434bfb2009-03-02 03:24:11 +00001026<a name="codegen">Target Independent Code Generator Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001027</h3>
Chris Lattner7795ea92008-06-08 02:45:07 +00001028
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001029<div>
Chris Lattner7795ea92008-06-08 02:45:07 +00001030
Mikhail Glushenkov25422542009-03-01 18:09:47 +00001031<p>We have put a significant amount of work into the code generator
Bill Wendling16133782011-10-26 18:46:16 +00001032 infrastructure, which allows us to implement more aggressive algorithms and
1033 make it run faster:</p>
Chris Lattner7795ea92008-06-08 02:45:07 +00001034
1035<ul>
Chad Rosier06da55e2011-05-27 22:50:46 +00001036<!--
1037<li></li>
1038-->
Chris Lattner0a1fd102007-09-21 03:54:09 +00001039</ul>
Chris Lattner0a1fd102007-09-21 03:54:09 +00001040</div>
1041
Chris Lattnerd1094e02009-03-02 02:37:32 +00001042<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001043<h3>
Chris Lattnerd1094e02009-03-02 02:37:32 +00001044<a name="x86">X86-32 and X86-64 Target Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001045</h3>
Chris Lattnerd1094e02009-03-02 02:37:32 +00001046
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001047<div>
Bill Wendling16133782011-10-26 18:46:16 +00001048
1049<p>New features and major changes in the X86 target include:</p>
Chris Lattnerd1094e02009-03-02 02:37:32 +00001050
1051<ul>
Benjamin Kramer8c8486d2011-11-26 11:14:54 +00001052 <li>The X86 backend, assembler and disassembler now completely support AVX.
1053 To enable it pass <code>-mavx</code> to the compiler.</li>
Bill Wendling16133782011-10-26 18:46:16 +00001054
Chris Lattnerdfcdd0c2011-11-15 22:48:24 +00001055 <li>The X86 backend now supports
1056 all <a href="http://llvm.org/PR879">inline assembly that uses the X86
1057 floating point stack</a>.</li>
1058
Bill Wendling16133782011-10-26 18:46:16 +00001059 <li>The CRC32 intrinsics have been renamed. The intrinsics were previously
1060 <code>@llvm.x86.sse42.crc32.[8|16|32]</code>
1061 and <code>@llvm.x86.sse42.crc64.[8|64]</code>. They have been renamed to
1062 <code>@llvm.x86.sse42.crc32.32.[8|16|32]</code> and
1063 <code>@llvm.x86.sse42.crc32.64.[8|64]</code>.</li>
Chad Rosierd1db4f82011-05-27 20:13:10 +00001064
Chris Lattnerd1094e02009-03-02 02:37:32 +00001065</ul>
1066
Chris Lattnerd1094e02009-03-02 02:37:32 +00001067</div>
Chris Lattner0a1fd102007-09-21 03:54:09 +00001068
1069<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001070<h3>
Chris Lattneraa61f412009-10-13 17:48:04 +00001071<a name="ARM">ARM Target Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001072</h3>
Chris Lattnerc92d7692009-03-01 02:30:21 +00001073
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001074<div>
Bill Wendling16133782011-10-26 18:46:16 +00001075
1076<p>New features of the ARM target include:</p>
Chris Lattnerc92d7692009-03-01 02:30:21 +00001077
1078<ul>
Jim Grosbach651e2ee2011-11-24 00:49:21 +00001079 <li>Reworked Set Jump Long Jump EH Lowering,</li>
1080 <li>improved support for Cortex-M series processors, and</li>
1081 <li>beta quality integrated assembler support.</li>
Bob Wilsone44f2982010-09-13 17:39:35 +00001082</ul>
Chris Lattneraa61f412009-10-13 17:48:04 +00001083</div>
Chris Lattner212a0862011-11-15 22:23:46 +00001084
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001085
1086<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001087<h3>
Akira Hatanakab89a58d2011-11-15 21:33:05 +00001088<a name="MIPS">MIPS Target Improvements</a>
1089</h3>
1090
1091<div>
1092
1093<p>New features and major changes in the MIPS target include:</p>
1094
1095<ul>
1096 <li>Most MIPS32r1 and r2 instructions are now supported.</li>
1097 <li>LE/BE MIPS32r1/r2 has been tested extensively.</li>
1098 <li>O32 ABI has been fully tested.</li>
1099 <li>MIPS backend has migrated to using the MC infrastructure for assembly printing. Initial support for direct object code emission has been implemented too.</li>
1100 <li>Delay slot filler has been updated. Now it tries to fill delay slots with useful instructions instead of always filling them with NOPs.</li>
1101 <li>Support for old-style JIT is complete.</li>
1102 <li>Support for old architectures (MIPS1 and MIPS2) has been removed.</li>
1103 <li>Initial support for MIPS64 has been added.</li>
1104</ul>
1105</div>
Chris Lattner212a0862011-11-15 22:23:46 +00001106
1107<!--=========================================================================-->
1108<h3>
1109 <a name="PTX">PTX Target Improvements</a>
1110</h3>
1111
1112<div>
Akira Hatanakab89a58d2011-11-15 21:33:05 +00001113
Chris Lattner212a0862011-11-15 22:23:46 +00001114 <p>
1115 The PTX back-end is still experimental, but is fairly usable for compute kernels
1116 in LLVM 3.0. Most scalar arithmetic is implemented, as well as intrinsics to
1117 access the special PTX registers and sync instructions. The major missing
1118 pieces are texture/sampler support and some vector operations.</p>
1119
1120 <p>That said, the backend is already being used for domain-specific languages
1121 and works well with the <a href="http://www.pcc.me.uk/~peter/libclc/">libclc
1122 library</a> to supply OpenCL built-ins. With it, you can use Clang to compile
1123 OpenCL code into PTX and execute it by loading the resulting PTX as a binary
1124 blob using the nVidia OpenCL library. It has been tested with several OpenCL
1125 programs, including some from the nVidia GPU Computing SDK, and the performance
1126 is on par with the nVidia compiler.</p>
1127
1128</div>
1129
Akira Hatanakab89a58d2011-11-15 21:33:05 +00001130<!--=========================================================================-->
1131<h3>
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001132<a name="OtherTS">Other Target Specific Improvements</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001133</h3>
Chris Lattnerc92d7692009-03-01 02:30:21 +00001134
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001135<div>
Bill Wendling16133782011-10-26 18:46:16 +00001136
Chris Lattner6d3eeec2011-11-10 20:15:40 +00001137 <p>PPC32/ELF va_arg was implemented.</p>
1138 <p>PPC32 initial support for .o file writing was implemented.</p>
Wesley Peck1c29a832011-11-14 18:56:41 +00001139 <p>MicroBlaze scheduling itineraries were added that model the
1140 3-stage and the 5-stage pipeline architectures. The 3-stage
1141 pipeline model can be selected with <code>-mcpu=mblaze3</code>
1142 and the 5-stage pipeline model can be selected with
1143 <code>-mcpu=mblaze5</code>.</p>
Chris Lattner6d3eeec2011-11-10 20:15:40 +00001144
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001145<ul>
Chad Rosier06da55e2011-05-27 22:50:46 +00001146<!--
1147<li></li>
1148-->
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001149</ul>
Bill Wendling16133782011-10-26 18:46:16 +00001150
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001151</div>
Chris Lattner6cb64032008-06-05 08:02:49 +00001152
1153<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001154<h3>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001155<a name="changes">Major Changes and Removed Features</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001156</h3>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001157
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001158<div>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001159
Bill Wendling2d3138c2011-08-03 22:18:20 +00001160<p>If you're already an LLVM user or developer with out-of-tree changes based on
1161 LLVM 2.9, this section lists some "gotchas" that you may run into upgrading
1162 from the previous release.</p>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001163
1164<ul>
Chris Lattner3dcdc292011-11-27 08:32:32 +00001165 <li>The <code>LLVMC</code> meta compiler driver was removed.</li>
Jay Foad2755e072011-08-04 10:43:43 +00001166 <li>The <code>LowerSetJmp</code> pass wasn't used effectively by any
1167 target and has been removed.</li>
Rafael Espindolaa45c20b2011-08-30 23:03:45 +00001168 <li>The old <code>TailDup</code> pass was not used in the standard pipeline
1169 and was unable to update ssa form, so it has been removed.
Eli Friedman02e737b2011-08-12 22:50:01 +00001170 <li>The syntax of volatile loads and stores in IR has been changed to
1171 "<code>load volatile</code>"/"<code>store volatile</code>". The old
1172 syntax ("<code>volatile load</code>"/"<code>volatile store</code>")
1173 is still accepted, but is now considered deprecated.</li>
Benjamin Kramera02af612011-11-25 21:26:00 +00001174 <li>The old atomic intrinsics (<code>llvm.memory.barrier</code> and
Eli Friedman2c25d9c2011-10-26 00:55:23 +00001175 <code>llvm.atomic.*</code>) are now gone. Please use the new atomic
1176 instructions, described in the <a href="Atomics.html">atomics guide</a>.
Devang Pateldbf83832008-10-14 20:03:43 +00001177</ul>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001178
NAKAMURA Takumi2b462b52011-08-22 23:22:05 +00001179<h4>Windows (32-bit)</h4>
1180<div>
Bill Wendling16133782011-10-26 18:46:16 +00001181
NAKAMURA Takumi2b462b52011-08-22 23:22:05 +00001182<ul>
1183 <li>On Win32(MinGW32 and MSVC), Windows 2000 will not be supported.
1184 Windows XP or higher is required.</li>
1185</ul>
Bill Wendling16133782011-10-26 18:46:16 +00001186
NAKAMURA Takumi2b462b52011-08-22 23:22:05 +00001187</div>
1188
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001189</div>
1190
Daniel Dunbarf70898a2010-10-04 20:11:41 +00001191<!--=========================================================================-->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001192<h3>
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001193<a name="api_changes">Internal API Changes</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001194</h3>
Daniel Dunbarf70898a2010-10-04 20:11:41 +00001195
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001196<div>
Daniel Dunbarf70898a2010-10-04 20:11:41 +00001197
Chris Lattnerc3a2c982011-04-06 00:45:11 +00001198<p>In addition, many APIs have changed in this release. Some of the major
Bill Wendling46ffaa92011-08-02 06:20:17 +00001199 LLVM API changes are:</p>
Daniel Dunbarf70898a2010-10-04 20:11:41 +00001200
1201<ul>
Benjamin Kramera02af612011-11-25 21:26:00 +00001202 <li>The biggest and most pervasive change is that llvm::Types are no longer
Bill Wendling16133782011-10-26 18:46:16 +00001203 returned or accepted as 'const' values. Instead, just pass around
Benjamin Kramera02af612011-11-25 21:26:00 +00001204 non-const Types.</li>
Chris Lattner14b3b4d2011-07-18 04:56:02 +00001205
Bill Wendling16133782011-10-26 18:46:16 +00001206 <li><code>PHINode::reserveOperandSpace</code> has been removed. Instead, you
1207 must specify how many operands to reserve space for when you create the
1208 PHINode, by passing an extra argument
1209 into <code>PHINode::Create</code>.</li>
Jay Foade03f15a2011-07-14 09:19:05 +00001210
Bill Wendling16133782011-10-26 18:46:16 +00001211 <li>PHINodes no longer store their incoming BasicBlocks as operands. Instead,
1212 the list of incoming BasicBlocks is stored separately, and can be accessed
1213 with new functions <code>PHINode::block_begin</code>
1214 and <code>PHINode::block_end</code>.</li>
Jay Foade03f15a2011-07-14 09:19:05 +00001215
Bill Wendling16133782011-10-26 18:46:16 +00001216 <li>Various functions now take an <code>ArrayRef</code> instead of either a
1217 pair of pointers (or iterators) to the beginning and end of a range, or a
1218 pointer and a length. Others now return an <code>ArrayRef</code> instead
1219 of a reference to a <code>SmallVector</code>
1220 or <code>std::vector</code>. These include:
Jay Foade03f15a2011-07-14 09:19:05 +00001221<ul>
1222<!-- Please keep this list sorted. -->
Jay Foad5bd375a2011-07-15 08:37:34 +00001223<li><code>CallInst::Create</code></li>
Jay Foade03f15a2011-07-14 09:19:05 +00001224<li><code>ComputeLinearIndex</code> (in <code>llvm/CodeGen/Analysis.h</code>)</li>
1225<li><code>ConstantArray::get</code></li>
1226<li><code>ConstantExpr::getExtractElement</code></li>
Jay Foaded8db7d2011-07-21 14:31:17 +00001227<li><code>ConstantExpr::getGetElementPtr</code></li>
1228<li><code>ConstantExpr::getInBoundsGetElementPtr</code></li>
Jay Foade03f15a2011-07-14 09:19:05 +00001229<li><code>ConstantExpr::getIndices</code></li>
1230<li><code>ConstantExpr::getInsertElement</code></li>
1231<li><code>ConstantExpr::getWithOperands</code></li>
Jay Foadf4b14a22011-07-19 13:32:40 +00001232<li><code>ConstantFoldCall</code> (in <code>llvm/Analysis/ConstantFolding.h</code>)</li>
1233<li><code>ConstantFoldInstOperands</code> (in <code>llvm/Analysis/ConstantFolding.h</code>)</li>
Jay Foade03f15a2011-07-14 09:19:05 +00001234<li><code>ConstantVector::get</code></li>
1235<li><code>DIBuilder::createComplexVariable</code></li>
1236<li><code>DIBuilder::getOrCreateArray</code></li>
1237<li><code>ExtractValueInst::Create</code></li>
1238<li><code>ExtractValueInst::getIndexedType</code></li>
1239<li><code>ExtractValueInst::getIndices</code></li>
1240<li><code>FindInsertedValue</code> (in <code>llvm/Analysis/ValueTracking.h</code>)</li>
Jay Foad528beda2011-07-19 14:42:50 +00001241<li><code>gep_type_begin</code> (in <code>llvm/Support/GetElementPtrTypeIterator.h</code>)</li>
1242<li><code>gep_type_end</code> (in <code>llvm/Support/GetElementPtrTypeIterator.h</code>)</li>
Jay Foadd1b78492011-07-25 09:48:08 +00001243<li><code>GetElementPtrInst::Create</code></li>
1244<li><code>GetElementPtrInst::CreateInBounds</code></li>
1245<li><code>GetElementPtrInst::getIndexedType</code></li>
Jay Foad857a48a2011-07-21 14:42:51 +00001246<li><code>InsertValueInst::Create</code></li>
1247<li><code>InsertValueInst::getIndices</code></li>
1248<li><code>InvokeInst::Create</code></li>
Jay Foad5bd375a2011-07-15 08:37:34 +00001249<li><code>IRBuilder::CreateCall</code></li>
Jay Foade03f15a2011-07-14 09:19:05 +00001250<li><code>IRBuilder::CreateExtractValue</code></li>
Jay Foad040dd822011-07-22 08:16:57 +00001251<li><code>IRBuilder::CreateGEP</code></li>
1252<li><code>IRBuilder::CreateInBoundsGEP</code></li>
Jay Foade03f15a2011-07-14 09:19:05 +00001253<li><code>IRBuilder::CreateInsertValue</code></li>
Jay Foad5bd375a2011-07-15 08:37:34 +00001254<li><code>IRBuilder::CreateInvoke</code></li>
Jay Foade03f15a2011-07-14 09:19:05 +00001255<li><code>MDNode::get</code></li>
1256<li><code>MDNode::getIfExists</code></li>
1257<li><code>MDNode::getTemporary</code></li>
1258<li><code>MDNode::getWhenValsUnresolved</code></li>
Jay Foadb992a632011-07-19 15:07:52 +00001259<li><code>SimplifyGEPInst</code> (in <code>llvm/Analysis/InstructionSimplify.h</code>)</li>
Jay Foadbf904772011-07-19 14:01:37 +00001260<li><code>TargetData::getIndexedOffset</code></li>
Jay Foade03f15a2011-07-14 09:19:05 +00001261</ul></li>
1262
Bill Wendling16133782011-10-26 18:46:16 +00001263 <li>All forms of <code>StringMap::getOrCreateValue</code> have been remove
1264 except for the one which takes a <code>StringRef</code>.</li>
Jay Foade03f15a2011-07-14 09:19:05 +00001265
Bill Wendling16133782011-10-26 18:46:16 +00001266 <li>The <code>LLVMBuildUnwind</code> function from the C API was removed. The
1267 LLVM <code>unwind</code> instruction has been deprecated for a long time
1268 and isn't used by the current front-ends. So this was removed during the
1269 exception handling rewrite.</li>
Bill Wendling46ffaa92011-08-02 06:20:17 +00001270
Bill Wendling16133782011-10-26 18:46:16 +00001271 <li>The <code>LLVMAddLowerSetJmpPass</code> function from the C API was
1272 removed because the <code>LowerSetJmp</code> pass was removed.</li>
Bill Wendling2d3138c2011-08-03 22:18:20 +00001273
Bill Wendling16133782011-10-26 18:46:16 +00001274 <li>The <code>DIBuilder</code> interface used by front ends to encode
1275 debugging information in the LLVM IR now expects clients to
1276 use <code>DIBuilder::finalize()</code> at the end of translation unit to
1277 complete debugging information encoding.</li>
Devang Patel2b8acaf2011-08-15 23:00:00 +00001278
Bill Wendling16133782011-10-26 18:46:16 +00001279 <li>The way the type system works has been
1280 rewritten: <code>PATypeHolder</code> and <code>OpaqueType</code> are gone,
1281 and all APIs deal with <code>Type*</code> instead of <code>const
1282 Type*</code>. If you need to create recursive structures, then create a
1283 named structure, and use <code>setBody()</code> when all its elements are
1284 built. Type merging and refining is gone too: named structures are not
1285 merged with other structures, even if their layout is identical. (of
1286 course anonymous structures are still uniqued by layout).</li>
Torok Edwin52cac092011-09-30 13:07:52 +00001287
Bill Wendling16133782011-10-26 18:46:16 +00001288 <li>TargetSelect.h moved to Support/ from Target/</li>
Torok Edwin52cac092011-09-30 13:07:52 +00001289
Bill Wendling16133782011-10-26 18:46:16 +00001290 <li>UpgradeIntrinsicCall no longer upgrades pre-2.9 intrinsic calls (for
1291 example <code>llvm.memset.i32</code>).</li>
Torok Edwin52cac092011-09-30 13:07:52 +00001292
Bill Wendling16133782011-10-26 18:46:16 +00001293 <li>It is mandatory to initialize all out-of-tree passes too and their dependencies now with
1294 <code>INITIALIZE_PASS{BEGIN,END,}</code>
1295 and <code>INITIALIZE_{PASS,AG}_DEPENDENCY</code>.</li>
Torok Edwin52cac092011-09-30 13:07:52 +00001296
Bill Wendling16133782011-10-26 18:46:16 +00001297 <li>The interface for MemDepResult in MemoryDependenceAnalysis has been
1298 enhanced with new return types Unknown and NonFuncLocal, in addition to
1299 the existing types Clobber, Def, and NonLocal.</li>
Daniel Dunbarf70898a2010-10-04 20:11:41 +00001300</ul>
Bill Wendling16133782011-10-26 18:46:16 +00001301
Daniel Dunbarf70898a2010-10-04 20:11:41 +00001302</div>
Chris Lattner1e4d5bc2008-10-13 17:57:36 +00001303
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001304</div>
1305
Chris Lattner3d482502003-10-02 04:57:28 +00001306<!-- *********************************************************************** -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001307<h2>
Misha Brukman80731b92003-11-22 00:38:41 +00001308 <a name="knownproblems">Known Problems</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001309</h2>
Chris Lattner3d482502003-10-02 04:57:28 +00001310<!-- *********************************************************************** -->
1311
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001312<div>
Misha Brukman80731b92003-11-22 00:38:41 +00001313
Bill Wendling16133782011-10-26 18:46:16 +00001314<p>This section contains significant known problems with the LLVM system, listed
1315 by component. If you run into a problem, please check
1316 the <a href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
1317 there isn't already one.</p>
Chris Lattner3d482502003-10-02 04:57:28 +00001318
Chris Lattnerb911de42004-03-14 02:03:02 +00001319<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001320<h3>
Chris Lattnerb911de42004-03-14 02:03:02 +00001321 <a name="experimental">Experimental features included with this release</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001322</h3>
Chris Lattner3d482502003-10-02 04:57:28 +00001323
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001324<div>
Chris Lattnerb911de42004-03-14 02:03:02 +00001325
Misha Brukmanfa50a222004-05-12 21:46:05 +00001326<p>The following components of this LLVM release are either untested, known to
Bill Wendling16133782011-10-26 18:46:16 +00001327 be broken or unreliable, or are in early development. These components
1328 should not be relied on, and bugs should not be filed against them, but they
1329 may be useful to some people. In particular, if you would like to work on
1330 one of these components, please contact us on
1331 the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev
1332 list</a>.</p>
Chris Lattnerb911de42004-03-14 02:03:02 +00001333
1334<ul>
Bill Wendling16133782011-10-26 18:46:16 +00001335 <li>The Alpha, Blackfin, CellSPU, MicroBlaze, MSP430, MIPS, PTX, SystemZ and
1336 XCore backends are experimental.</li>
1337
1338 <li><tt>llc</tt> "<tt>-filetype=obj</tt>" is experimental on all targets other
1339 than darwin and ELF X86 systems.</li>
Chris Lattnerb911de42004-03-14 02:03:02 +00001340</ul>
1341
1342</div>
Chris Lattner3d482502003-10-02 04:57:28 +00001343
Misha Brukman80731b92003-11-22 00:38:41 +00001344<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001345<h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001346 <a name="x86-be">Known problems with the X86 back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001347</h3>
John Criswell3bdbd302005-11-08 21:11:33 +00001348
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001349<div>
John Criswell3bdbd302005-11-08 21:11:33 +00001350
1351<ul>
Chris Lattner0e439c9b2011-11-17 01:42:23 +00001352 <li>The X86-64 backend <a href="http://llvm.org/PR1740">does not yet support
1353 the <tt>va_arg</tt> LLVM IR instruction</a>. Currently, front-ends support
1354 variadic argument constructs on X86-64 by lowering them manually.</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001355</ul>
1356
1357</div>
1358
1359<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001360<h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001361 <a name="ppc-be">Known problems with the PowerPC back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001362</h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001363
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001364<div>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001365
1366<ul>
Roman Divacky2634d772011-10-30 07:49:04 +00001367 <li>The PPC32/ELF support lacks PIC support.</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001368</ul>
1369
1370</div>
1371
1372<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001373<h3>
Chris Lattner97beb512007-05-14 06:56:09 +00001374 <a name="arm-be">Known problems with the ARM back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001375</h3>
Chris Lattner97beb512007-05-14 06:56:09 +00001376
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001377<div>
Chris Lattner97beb512007-05-14 06:56:09 +00001378
1379<ul>
Bill Wendling16133782011-10-26 18:46:16 +00001380 <li>Thumb mode works only on ARMv6 or higher processors. On sub-ARMv6
1381 processors, thumb programs can crash or produce wrong results
1382 (<a href="http://llvm.org/PR1388">PR1388</a>).</li>
1383
1384 <li>Compilation for ARM Linux OABI (old ABI) is supported but not fully
1385 tested.</li>
Chris Lattner97beb512007-05-14 06:56:09 +00001386</ul>
1387
1388</div>
1389
1390<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001391<h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001392 <a name="sparc-be">Known problems with the SPARC back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001393</h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001394
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001395<div>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001396
1397<ul>
Bill Wendling16133782011-10-26 18:46:16 +00001398 <li>The SPARC backend only supports the 32-bit SPARC ABI (-m32); it does not
1399 support the 64-bit SPARC ABI (-m64).</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001400</ul>
1401
1402</div>
1403
1404<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001405<h3>
Bruno Cardoso Lopes24eb3de2008-10-25 14:56:26 +00001406 <a name="mips-be">Known problems with the MIPS back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001407</h3>
Bruno Cardoso Lopes24eb3de2008-10-25 14:56:26 +00001408
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001409<div>
Bruno Cardoso Lopes24eb3de2008-10-25 14:56:26 +00001410
1411<ul>
Bill Wendling16133782011-10-26 18:46:16 +00001412 <li>64-bit MIPS targets are not supported yet.</li>
Bruno Cardoso Lopes24eb3de2008-10-25 14:56:26 +00001413</ul>
1414
1415</div>
1416
1417<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001418<h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001419 <a name="alpha-be">Known problems with the Alpha back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001420</h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001421
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001422<div>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001423
1424<ul>
Bill Wendling16133782011-10-26 18:46:16 +00001425 <li>On 21164s, some rare FP arithmetic sequences which may trap do not have
1426 the appropriate nops inserted to ensure restartability.</li>
John Criswell3bdbd302005-11-08 21:11:33 +00001427</ul>
Bill Wendling16133782011-10-26 18:46:16 +00001428
John Criswell3bdbd302005-11-08 21:11:33 +00001429</div>
1430
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001431<!-- ======================================================================= -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001432<h3>
Chris Lattner97beb512007-05-14 06:56:09 +00001433 <a name="c-be">Known problems with the C back-end</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001434</h3>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001435
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001436<div>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001437
Chris Lattner086d2692010-09-29 05:34:42 +00001438<p>The C backend has numerous problems and is not being actively maintained.
Bill Wendling16133782011-10-26 18:46:16 +00001439 Depending on it for anything serious is not advised.</p>
Chris Lattner086d2692010-09-29 05:34:42 +00001440
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001441<ul>
Bill Wendling16133782011-10-26 18:46:16 +00001442 <li><a href="http://llvm.org/PR802">The C backend has only basic support for
1443 inline assembly code</a>.</li>
1444
1445 <li><a href="http://llvm.org/PR1658">The C backend violates the ABI of common
1446 C++ programs</a>, preventing intermixing between C++ compiled by the CBE
1447 and C++ code compiled with <tt>llc</tt> or native compilers.</li>
1448
1449 <li>The C backend does not support all exception handling constructs.</li>
1450
1451 <li>The C backend does not support arbitrary precision integers.</li>
Chris Lattnerb81f10e2006-11-18 07:51:14 +00001452</ul>
1453
1454</div>
John Criswell3bdbd302005-11-08 21:11:33 +00001455
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001456</div>
1457
Chris Lattner3d482502003-10-02 04:57:28 +00001458<!-- *********************************************************************** -->
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001459<h2>
Misha Brukman80731b92003-11-22 00:38:41 +00001460 <a name="additionalinfo">Additional Information</a>
NAKAMURA Takumi64835132011-04-18 01:17:51 +00001461</h2>
Chris Lattner3d482502003-10-02 04:57:28 +00001462<!-- *********************************************************************** -->
1463
NAKAMURA Takumi3ad28282011-04-21 01:52:00 +00001464<div>
Misha Brukman80731b92003-11-22 00:38:41 +00001465
Bill Wendling16133782011-10-26 18:46:16 +00001466<p>A wide variety of additional information is available on
1467 the <a href="http://llvm.org/">LLVM web page</a>, in particular in
1468 the <a href="http://llvm.org/docs/">documentation</a> section. The web page
1469 also contains versions of the API documentation which is up-to-date with the
1470 Subversion version of the source code. You can access versions of these
1471 documents specific to this release by going into the "<tt>llvm/doc/</tt>"
1472 directory in the LLVM tree.</p>
Chris Lattner3d482502003-10-02 04:57:28 +00001473
Misha Brukman80731b92003-11-22 00:38:41 +00001474<p>If you have any questions or comments about LLVM, please feel free to contact
Bill Wendling16133782011-10-26 18:46:16 +00001475 us via the <a href="http://llvm.org/docs/#maillist"> mailing lists</a>.</p>
Chris Lattner3d482502003-10-02 04:57:28 +00001476
Misha Brukman80731b92003-11-22 00:38:41 +00001477</div>
Chris Lattner3d482502003-10-02 04:57:28 +00001478
1479<!-- *********************************************************************** -->
Chris Lattner3d482502003-10-02 04:57:28 +00001480
Misha Brukman80731b92003-11-22 00:38:41 +00001481<hr>
Misha Brukman68aab3b2003-11-22 01:23:39 +00001482<address>
Misha Brukman8834af92003-12-21 22:53:21 +00001483 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +00001484 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
Misha Brukman68aab3b2003-11-22 01:23:39 +00001485 <a href="http://validator.w3.org/check/referer"><img
Misha Brukman86242e12008-12-11 17:34:48 +00001486 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
Misha Brukman68aab3b2003-11-22 01:23:39 +00001487
Chris Lattnere0c1df42007-05-18 00:44:29 +00001488 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Misha Brukman80731b92003-11-22 00:38:41 +00001489 Last modified: $Date$
Misha Brukman68aab3b2003-11-22 01:23:39 +00001490</address>
Chris Lattner3d482502003-10-02 04:57:28 +00001491
Misha Brukman80731b92003-11-22 00:38:41 +00001492</body>
1493</html>