Google Git
Sign in
llvm / llvm / refs/heads/release_32 / . / docs / ReleaseNotes.html
blob: 290fbf6908354b242a8ae8c7a94b387182572eb4 [file] [log] [blame]
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<link rel="stylesheet" href="_static/llvm.css" type="text/css">
<title>LLVM 3.2 Release Notes</title>
</head>
<body>
<h1>LLVM 3.2 Release Notes</h1>
<div>
<img style="float:right" src="http://llvm.org/img/DragonSmall.png"
width="136" height="136" alt="LLVM Dragon Logo">
</div>
<ol>
<li><a href="#intro">Introduction</a></li>
<li><a href="#subproj">Sub-project Status Update</a></li>
<li><a href="#externalproj">External Projects Using LLVM 3.2</a></li>
<li><a href="#whatsnew">What's New in LLVM?</a></li>
<li><a href="GettingStarted.html">Installation Instructions</a></li>
<li><a href="#knownproblems">Known Problems</a></li>
<li><a href="#additionalinfo">Additional Information</a></li>
</ol>
<div class="doc_author">
<p>Written by the <a href="http://llvm.org/">LLVM Team</a></p>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="intro">Introduction</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>This document contains the release notes for the LLVM Compiler
Infrastructure, release 3.2. Here we describe the status of LLVM, including
major improvements from the previous release, improvements in various
sub-projects of LLVM, and some of the current users of the code. All LLVM
releases may be downloaded from the <a href="http://llvm.org/releases/">LLVM
releases web site</a>.</p>
<p>For more information about LLVM, including information about the latest
release, please check out the <a href="http://llvm.org/">main LLVM web
site</a>. If you have questions or comments,
the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM
Developer's Mailing List</a> is a good place to send them.</p>
<p>Note that if you are reading this file from a Subversion checkout or the main
LLVM web page, this document applies to the <i>next</i> release, not the
current one. To see the release notes for a specific release, please see the
<a href="http://llvm.org/releases/">releases page</a>.</p>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="subproj">Sub-project Status Update</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>The LLVM 3.2 distribution currently consists of production-quality code
from the core LLVM repository, which roughly includes the LLVM optimizers,
code generators and supporting tools, as well as Clang, DragonEgg and
compiler-rt sub-project repositories. In addition to this code, the LLVM
Project includes other sub-projects that are in development. Here we
include updates on these sub-projects.</p>
<!--=========================================================================-->
<h3>
<a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
</h3>
<div>
<p><a href="http://clang.llvm.org/">Clang</a> is an LLVM front end for the C,
C++, and Objective-C languages. Clang aims to provide a better user
experience through expressive diagnostics, a high level of conformance to
language standards, fast compilation, and low memory use. Like LLVM, Clang
provides a modular, library-based architecture that makes it suitable for
creating or integrating with other development tools.</p>
<p>In the LLVM 3.2 time-frame, the Clang team has made many improvements.
Highlights include:</p>
<ul>
<li>Improvements to Clang's diagnostics</li>
<li>Support for tls_model attribute</li>
<li>Type safety attributes</li>
</ul>
<p>For more details about the changes to Clang since the 3.1 release, see the
<a href="http://llvm.org/releases/3.2/tools/clang/docs/ReleaseNotes.html">Clang 3.2 release
notes.</a></p>
<p>If Clang rejects your code but another compiler accepts it, please take a
look at the <a href="http://clang.llvm.org/compatibility.html">language
compatibility</a> guide to make sure this is not intentional or a known
issue.</p>
</div>
<!--=========================================================================-->
<h3>
<a name="dragonegg">DragonEgg: GCC front-ends, LLVM back-end</a>
</h3>
<div>
<p><a href="http://dragonegg.llvm.org/">DragonEgg</a> is a
<a href="http://gcc.gnu.org/wiki/plugins">gcc plugin</a> that replaces GCC's
optimizers and code generators with LLVM's. It works with gcc-4.5 and gcc-4.6
(and partially with gcc-4.7), can target the x86-32/x86-64 and ARM processor
families, and has been successfully used on the Darwin, FreeBSD, KFreeBSD,
Linux and OpenBSD platforms. It fully supports Ada, C, C++ and Fortran. It
has partial support for Go, Java, Obj-C and Obj-C++.</p>
<p>The 3.2 release has the following notable changes:</p>
<ul>
<li>Able to load LLVM plugins such as Polly.</li>
<li>Supports thread-local storage models.</li>
<li>Passes knowledge of variable lifetimes to the LLVM optimizers.</li>
<li>No longer requires GCC to be built with LTO support.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="compiler-rt">compiler-rt: Compiler Runtime Library</a>
</h3>
<div>
<p>The LLVM <a href="http://compiler-rt.llvm.org/">compiler-rt project</a>
is a simple library that provides an implementation of the low-level
target-specific hooks required by code generation and other runtime
components. For example, when compiling for a 32-bit target, converting a
double to a 64-bit unsigned integer is compiled into a runtime call to the
<code>__fixunsdfdi</code> function. The compiler-rt library provides highly
optimized implementations of this and other low-level routines (some are 3x
faster than the equivalent libgcc routines).</p>
<p>The 3.2 release has the following notable changes:</p>
<ul>
<li><a href="http://llvm.org/releases/3.2/tools/clang/docs/ThreadSanitizer.html">ThreadSanitizer (TSan)</a> - data race detector run-time library for C/C++ has been added.</li>
<li>Improvements to <a href="http://llvm.org/releases/3.2/tools/clang/docs/AddressSanitizer.html">AddressSanitizer</a> including: better portability
(OSX, Android NDK), support for cmake based builds, enhanced error reporting and lots of bug fixes.</li>
<li>Added support for A6 'Swift' CPU.</li>
<li><code>divsi3</code> function has been enhanced to take advantage of a hardware unsigned divide when it is available.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="lldb">LLDB: Low Level Debugger</a>
</h3>
<div>
<p><a href="http://lldb.llvm.org">LLDB</a> is a ground-up implementation of a
command line debugger, as well as a debugger API that can be used from other
applications. LLDB makes use of the Clang parser to provide high-fidelity
expression parsing (particularly for C++) and uses the LLVM JIT for target
support.</p>
<p>The 3.2 release has the following notable changes:</p>
<ul>
<li>Linux build fixes for clang (see <a href="http://lldb.llvm.org/build.html">Building LLDB</a>)</li>
<li>Some Linux stability and usability improvements</li>
<li>Switch expression evaluation to use MCJIT (from legacy JIT) on Linux</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="libc++">libc++: C++ Standard Library</a>
</h3>
<div>
<p>Like compiler_rt, libc++ is now <a href="DeveloperPolicy.html#license">dual
licensed</a> under the MIT and UIUC license, allowing it to be used more
permissively.</p>
<p>Within the LLVM 3.2 time-frame there were the following highlights:</p>
<ul>
<li> C++11 shared_ptr atomic access API (20.7.2.5) has been implemented.</li>
<li>Applied noexcept and constexpr throughout library.</li>
<li>Improved C++11 conformance in associative container emplace.</li>
<li>Performance improvements in: std::rotate algorithm and I/O.</li>
<li>Operator new/delete and type_infos for exception types moved from libc++ to libc++abi.</li>
<li>Bug fixes in: <code>&lt;atomic&gt;</code>; vector<code>&lt;bool&gt;</code> algorithms,
<code>&lt;future&gt;</code>,<code>&lt;tuple&gt;</code>,
<code>&lt;type_traits&gt;</code>,<code>&lt;fstream&gt;</code>,<code>&lt;istream&gt;</code>,
<code>&lt;iterator&gt;</code>, <code>&lt;condition_variable&gt;</code>,<code>&lt;complex&gt;</code> as well as visibility fixes.
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="vmkit">VMKit</a>
</h3>
<div>
<p>The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation
of a Java Virtual Machine (Java VM or JVM) that uses LLVM for static and
just-in-time compilation.</p>
<p>The 3.2 release has the following notable changes:</p>
<ul>
<li>Bug fixes only, no functional changes.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="Polly">Polly: Polyhedral Optimizer</a>
</h3>
<div>
<p><a href="http://polly.llvm.org/">Polly</a> is an <em>experimental</em>
optimizer for data locality and parallelism. It currently provides high-level
loop optimizations and automatic parallelization (using the OpenMP run time).
Work in the area of automatic SIMD and accelerator code generation was
started.</p>
<p>Within the LLVM 3.2 time-frame there were the following highlights:</p>
<ul>
<li>isl, the integer set library used by Polly, was relicensed under the MIT license.</li>
<li>isl based code generation.</li>
<li>MIT licensed replacement for CLooG (LGPLv2).</li>
<li>Fine grained option handling (separation of core and border computations, control overhead vs. code size).</li>
<li>Support for FORTRAN and Dragonegg.</li>
<li>OpenMP code generation fixes.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="StaticAnalyzer">Clang Static Analyzer</a>
</h3>
<div>
<p>The <a href="http://clang-analyzer.llvm.org/">Clang Static Analyzer</a>
is an advanced source code analysis tool integrated into Clang that performs
a deep analysis of code to find potential bugs.</p>
<p>In the LLVM 3.2 release, the static analyzer has made significant improvements
in many areas, with notable highlights such as:</p>
<ul>
<li>Improved interprocedural analysis within a translation unit (see details below), which greatly amplified the analyzer's ability to find bugs.</li>
<li>New infrastructure to model &quot;well-known&quot; APIs, allowing the analyzer to do a much better job when modeling calls to such functions.</li>
<li>Significant improvements to the APIs to write static analyzer checkers, with a more unified way of representing function/method calls in the checker API. Details can be found in the <a href="http://llvm.org/devmtg/2012-11#talk13">Building a Checker in 24 hours</a> talk.
</ul>
<p>The release specifically includes notable improvements for Objective-C analysis, including:</p>
<ul>
<li>Interprocedural analysis for Objective-C methods.</li>
<li>Interprocedural analysis of calls to &quot;blocks&quot;.</li>
<li>Precise modeling of GCD APIs such as <tt>dispatch_once</tt> and friends.</li>
<li>Improved support for recently added Objective-C constructs such as array and dictionary literals.</li>
</ul>
<p>The release specifically includes notable improvements for C++ analysis, including:</p>
<ul>
<li>Interprocedural analysis for C++ methods (within a translation unit).</li>
<li>More precise modeling of C++ initializers and destructors.</li>
</ul>
<p>Finally, this release includes many small improvements to <tt>scan-build</tt>, which can be used to drive the analyzer from the command line or a continuous integration system. This includes a directory-traversal issue, which could cause potential security problems in some cases. We would like to acknowledge Tim Brown of Portcullis Computer Security Ltd for reporting this issue.</p>
</div>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="externalproj">External Open Source Projects Using LLVM 3.2</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>An exciting aspect of LLVM is that it is used as an enabling technology for
a lot of other language and tools projects. This section lists some of the
projects that have already been updated to work with LLVM 3.2.</p>
<h3>Crack</h3>
<div>
<p><a href="http://code.google.com/p/crack-language/">Crack</a> aims to provide
the ease of development of a scripting language with the performance of a
compiled language. The language derives concepts from C++, Java and Python,
incorporating object-oriented programming, operator overloading and strong
typing.</p>
</div>
<h3>EmbToolkit</h3>
<div>
<p><a href="http://www.embtoolkit.org/">EmbToolkit</a> provides Linux cross-compiler
toolchain/SDK (GCC/binutils/C library (uclibc,eglibc,musl)), a build system for
package cross-compilation and optionally various root file systems.
It supports ARM and MIPS. There is an ongoing effort to provide a clang+llvm
environment for the 3.2 releases,
</p>
</div>
<h3>FAUST</h3>
<div>
<p><a href="http://faust.grame.fr/">FAUST</a> is a compiled language for
real-time audio signal processing. The name FAUST stands for Functional
AUdio STream. Its programming model combines two approaches: functional
programming and block diagram composition. In addition with the C, C++, Java,
JavaScript output formats, the Faust compiler can generate LLVM bitcode, and
works with LLVM 2.7-3.2.</p>
</div>
<h3>Glasgow Haskell Compiler (GHC)</h3>
<div>
<p><a href="http://www.haskell.org/ghc/">GHC</a> is an open source compiler and
programming suite for Haskell, a lazy functional programming language. It
includes an optimizing static compiler generating good code for a variety of
platforms, together with an interactive system for convenient, quick
development.</p>
<p>GHC 7.0 and onwards include an LLVM code generator, supporting LLVM 2.8 and
later.</p>
</div>
<h3>Julia</h3>
<div>
<p><a href="https://github.com/JuliaLang/julia">Julia</a> is a high-level,
high-performance dynamic language for technical computing. It provides a
sophisticated compiler, distributed parallel execution, numerical accuracy,
and an extensive mathematical function library. The compiler uses type
inference to generate fast code without any type declarations, and uses
LLVM's optimization passes and JIT compiler. The
<a href="http://julialang.org/"> Julia Language</a> is designed
around multiple dispatch, giving programs a large degree of flexibility. It
is ready for use on many kinds of problems.</p>
</div>
<h3>LLVM D Compiler</h3>
<div>
<p><a href="https://github.com/ldc-developers/ldc">LLVM D Compiler</a> (LDC) is
a compiler for the D programming Language. It is based on the DMD frontend
and uses LLVM as backend.</p>
</div>
<h3>Open Shading Language</h3>
<div>
<p><a href="https://github.com/imageworks/OpenShadingLanguage/">Open Shading
Language (OSL)</a> is a small but rich language for programmable shading in
advanced global illumination renderers and other applications, ideal for
describing materials, lights, displacement, and pattern generation. It uses
LLVM to JIT complex shader networks to x86 code at runtime.</p>
<p>OSL was developed by Sony Pictures Imageworks for use in its in-house
renderer used for feature film animation and visual effects, and is
distributed as open source software with the "New BSD" license.
It has been used for all the shading on such films as The Amazing Spider-Man,
Men in Black III, Hotel Transylvania, and may other films in-progress,
and also has been incorporated into several commercial and open source
rendering products such as Blender, VRay, and Autodesk Beast.</p>
</div>
<h3>Portable OpenCL (pocl)</h3>
<div>
<p>In addition to producing an easily portable open source OpenCL
implementation, another major goal of <a href="http://pocl.sourceforge.net/">
pocl</a> is improving performance portability of OpenCL programs with
compiler optimizations, reducing the need for target-dependent manual
optimizations. An important part of pocl is a set of LLVM passes used to
statically parallelize multiple work-items with the kernel compiler, even in
the presence of work-group barriers. This enables static parallelization of
the fine-grained static concurrency in the work groups in multiple ways
(SIMD, VLIW, superscalar,...).</p>
</div>
<h3>Pure</h3>
<div>
<p><a href="http://pure-lang.googlecode.com/">Pure</a> is an
algebraic/functional programming language based on term rewriting. Programs
are collections of equations which are used to evaluate expressions in a
symbolic fashion. The interpreter uses LLVM as a backend to JIT-compile Pure
programs to fast native code. Pure offers dynamic typing, eager and lazy
evaluation, lexical closures, a hygienic macro system (also based on term
rewriting), built-in list and matrix support (including list and matrix
comprehensions) and an easy-to-use interface to C and other programming
languages (including the ability to load LLVM bitcode modules, and inline C,
C++, Fortran and Faust code in Pure programs if the corresponding
LLVM-enabled compilers are installed).</p>
<p>Pure version 0.56 has been tested and is known to work with LLVM 3.2 (and
continues to work with older LLVM releases >= 2.5).</p>
</div>
<h3>TTA-based Co-design Environment (TCE)</h3>
<div>
<p><a href="http://tce.cs.tut.fi/">TCE</a> is a toolset for designing
application-specific processors (ASP) based on the Transport triggered
architecture (TTA). The toolset provides a complete co-design flow from C/C++
programs down to synthesizable VHDL/Verilog and parallel program binaries.
Processor customization points include the register files, function units,
supported operations, and the interconnection network.</p>
<p>TCE uses Clang and LLVM for C/C++ language support, target independent
optimizations and also for parts of code generation. It generates new
LLVM-based code generators "on the fly" for the designed TTA processors and
loads them in to the compiler backend as runtime libraries to avoid
per-target recompilation of larger parts of the compiler chain.</p>
</div>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="whatsnew">What's New in LLVM 3.2?</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>This release includes a huge number of bug fixes, performance tweaks and
minor improvements. Some of the major improvements and new features are
listed in this section.</p>
<!--=========================================================================-->
<h3>
<a name="majorfeatures">Major New Features</a>
</h3>
<div>
<!-- Features that need text if they're finished for 3.2:
ARM EHABI
combiner-aa?
strong phi elim
loop dependence analysis
CorrelatedValuePropagation
lib/Transforms/IPO/MergeFunctions.cpp => consider for 3.2.
Integrated assembler on by default for arm/thumb?
-->
<!-- Near dead:
Analysis/RegionInfo.h + Dom Frontiers
SparseBitVector: used in LiveVar.
llvm/lib/Archive - replace with lib object?
-->
<p>LLVM 3.2 includes several major changes and big features:</p>
<ul>
<li>Loop Vectorizer.</li>
<li>New implementation of SROA.</li>
<li>New NVPTX back-end (replacing existing PTX back-end) based on NVIDIA sources.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="coreimprovements">LLVM IR and Core Improvements</a>
</h3>
<div>
<p>LLVM IR has several new features for better support of new targets and that
expose new optimization opportunities:</p>
<ul>
<li>Thread local variables may have a specified TLS model. See the
<a href="LangRef.html#globalvars">Language Reference Manual</a>.</li>
<li>'TYPE_CODE_FUNCTION_OLD' type code and autoupgrade code for old function attributes format has been removed.</li>
<li>Internal representation of the Attributes class has been converted into a pointer to an
opaque object that's uniqued by and stored in the LLVMContext object.
The Attributes class then becomes a thin wrapper around this opaque object.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="optimizer">Optimizer Improvements</a>
</h3>
<div>
<p>In addition to many minor performance tweaks and bug fixes, this release
includes a few major enhancements and additions to the optimizers:</p>
<p> Loop Vectorizer - We've added a loop vectorizer and we are now able to
vectorize small loops. The loop vectorizer is disabled by default and
can be enabled using the <b>-mllvm -vectorize-loops</b> flag.
The SIMD vector width can be specified using the flag
<b>-mllvm -force-vector-width=4</b>.
The default value is <b>0</b> which means auto-select.
<br/>
We can now vectorize this function:
<pre class="doc_code">
unsigned sum_arrays(int *A, int *B, int start, int end) {
unsigned sum = 0;
for (int i = start; i &lt; end; ++i)
sum += A[i] + B[i] + i;
return sum;
}
</pre>
We vectorize under the following loops:
<ul>
<li>The inner most loops must have a single basic block.</li>
<li>The number of iterations are known before the loop starts to execute.</li>
<li>The loop counter needs to be incremented by one.</li>
<li>The loop trip count <b>can</b> be a variable.</li>
<li>Loops do <b>not</b> need to start at zero.</li>
<li>The induction variable can be used inside the loop.</li>
<li>Loop reductions are supported.</li>
<li>Arrays with affine access pattern do <b>not</b> need to be marked as 'noalias' and are checked at runtime.</li>
</ul>
</p>
<p>SROA - We&#8217;ve re-written SROA to be significantly more powerful and generate
code which is much more friendly to the rest of the optimization pipeline.
Previously this pass had scaling problems that required it to only operate on
relatively small aggregates, and at times it would mistakenly replace a large
aggregate with a single very large integer in order to make it a scalar SSA
value. The result was a large number of i1024 and i2048 values representing any
small stack buffer. These in turn slowed down many subsequent optimization
paths.</p>
<p>The new SROA pass uses a different algorithm that allows it to only promote to
scalars the pieces of the aggregate actively in use. Because of this it doesn&#8217;t
require any thresholds. It also always deduces the scalar values from the uses
of the aggregate rather than the specific LLVM type of the aggregate. These
features combine to both optimize more code with the pass but to improve the
compile time of many functions dramatically.</p>
<ul>
<li>Branch weight metadata is preserved through more of the optimizer.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="mc">MC Level Improvements</a>
</h3>
<div>
<p>The LLVM Machine Code (aka MC) subsystem was created to solve a number of
problems in the realm of assembly, disassembly, object file format handling,
and a number of other related areas that CPU instruction-set level tools work
in. For more information, please see the
<a href="http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html">Intro
to the LLVM MC Project Blog Post</a>.</p>
<ul>
<li> Added support for following assembler directives: <code>.ifb</code>, <code>.ifnb</code>, <code>.ifc</code>,
<code>.ifnc</code>, <code>.purgem</code>, <code>.rept</code> and <code>.version</code> (ELF) as well as Darwin specific
<code>.pushsection</code>, <code>.popsection</code> and <code>.previous</code> .</li>
<li>Enhanced handling of <code>.lcomm directive</code>.</li>
<li>MS style inline assembler: added implementation of the offset and TYPE operators.</li>
<li>Targets can specify minimum supported NOP size for NOP padding.</li>
<li>ELF improvements: added support for generating ELF objects on Windows.</li>
<li>MachO improvements: symbol-difference variables are marked as N_ABS, added direct-to-object attribute for data-in-code markers.</li>
<li>Added support for annotated disassembly output for x86 and arm targets.</li>
<li>Arm support has been improved by adding support for ARM TARGET2 relocation
and fixing hadling of ARM-style "$d.*" labels.</li>
<li>Implemented local-exec TLS on PowerPC.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="codegen">Target Independent Code Generator Improvements</a>
</h3>
<div>
<p>Stack Coloring - We have implemented a new optimization pass
to merge stack objects which are used in disjoin areas of the code.
This optimization reduces the required stack space significantly, in cases
where it is clear to the optimizer that the stack slot is not shared.
We use the lifetime markers to tell the codegen that a certain alloca
is used within a region.</p>
<p> We now merge consecutive loads and stores. </p>
<p>We have put a significant amount of work into the code generator
infrastructure, which allows us to implement more aggressive algorithms and
make it run faster:</p>
<p> We added new TableGen infrastructure to support bundling for
Very Long Instruction Word (VLIW) architectures. TableGen can now
automatically generate a deterministic finite automaton from a VLIW
target's schedule description which can be queried to determine
legal groupings of instructions in a bundle.</p>
<p> We have added a new target independent VLIW packetizer based on the
DFA infrastructure to group machine instructions into bundles.</p>
<p> We have added new TableGen infrastructure to support relationship maps
between instructions. This feature enables TableGen to automatically
construct a set of relation tables and query functions that can be used
to switch between various forms of instructions. For more information,
please refer to <a href="http://llvm.org/docs/HowToUseInstrMappings.html">
How To Use Instruction Mappings</a>.</p>
</div>
<h4>
<a name="blockplacement">Basic Block Placement</a>
</h4>
<div>
<p>A probability based block placement and code layout algorithm was added to
LLVM's code generator. This layout pass supports probabilities derived from
static heuristics as well as source code annotations such as
<code>__builtin_expect</code>.</p>
</div>
<!--=========================================================================-->
<h3>
<a name="x86">X86-32 and X86-64 Target Improvements</a>
</h3>
<div>
<p>New features and major changes in the X86 target include:</p>
<ul>
<li>Small codegen optimizations, especially for AVX2.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="ARM">ARM Target Improvements</a>
</h3>
<div>
<p>New features of the ARM target include:</p>
<ul>
<li>Support and performance tuning for the A6 'Swift' CPU.</li>
</ul>
<!--_________________________________________________________________________-->
<h4>
<a name="armintegratedassembler">ARM Integrated Assembler</a>
</h4>
<div>
<p>The ARM target now includes a full featured macro assembler, including
direct-to-object module support for clang. The assembler is currently enabled
by default for Darwin only pending testing and any additional necessary
platform specific support for Linux.</p>
<p>Full support is included for Thumb1, Thumb2 and ARM modes, along with
sub-target and CPU specific extensions for VFP2, VFP3 and NEON.</p>
<p>The assembler is Unified Syntax only (see ARM Architecural Reference Manual
for details). While there is some, and growing, support for pre-unfied
(divided) syntax, there are still significant gaps in that support.</p>
</div>
</div>
<!--=========================================================================-->
<h3>
<a name="MIPS">MIPS Target Improvements</a>
</h3>
<div>
<p>New features and major changes in the MIPS target include:</p>
<ul>
<li>Integrated assembler support:
MIPS32 works for both PIC and static, known limitation is the PR14456 where
R_MIPS_GPREL16 relocation is generated with the wrong addend.
MIPS64 support is incomplete, for example exception handling is not working.</li>
<li>Support for fast calling convention has been added.</li>
<li>Support for Android MIPS toolchain has been added to clang driver.</li>
<li>Added clang driver support for MIPS N32 ABI through "-mabi=n32" option.</li>
<li>MIPS32 and MIPS64 disassembler has been implemented.</li>
<li>Support for compiling programs with large GOTs (exceeding 64kB in size) has been added
through llc option "-mxgot".</li>
<li>Added experimental support for MIPS32 DSP intrinsics.</li>
<li>Experimental support for MIPS16 with following limitations: only soft float is supported,
C++ exceptions are not supported, large stack frames (> 32000 bytes) are not supported,
direct object code emission is not supported only .s .</li>
<li>Standalone assembler (llvm-mc): implementation is in progress and considered experimental.</li>
<li>All classic JIT and MCJIT tests pass on Little and Big Endian MIPS32 platforms.</li>
<li>Inline asm support: all common constraints and operand modifiers have been implemented.</li>
<li>Added tail call optimization support, use llc option "-enable-mips-tail-calls"
or clang options "-mllvm -enable-mips-tail-calls"to enable it.</li>
<li>Improved register allocation by removing registers $fp, $gp, $ra and $at from the list of reserved registers.</li>
<li>Long branch expansion pass has been implemented, which expands branch
instructions with offsets that do not fit in the 16-bit field.</li>
<li>Cavium Octeon II board is used for testing builds (llvm-mips-linux builder).</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="PowerPC">PowerPC Target Improvements</a>
</h3>
<div>
<p>Many fixes and changes across LLVM (and Clang) for better compliance with
the 64-bit PowerPC ELF Application Binary Interface, interoperability with
GCC, and overall 64-bit PowerPC support. Some highlights include:</p>
<ul>
<li> MCJIT support added.</li>
<li> PPC64 relocation support and (small code model) TOC handling
added.</li>
<li> Parameter passing and return value fixes (alignment issues,
padding, varargs support, proper register usage, odd-sized
structure support, float support, extension of return values
for i32 return values).</li>
<li> Fixes in spill and reload code for vector registers.</li>
<li> C++ exception handling enabled.</li>
<li> Changes to remediate double-rounding compatibility issues with
respect to GCC behavior.</li>
<li> Refactoring to disentangle ppc64-elf-linux ABI from Darwin
ppc64 ABI support.</li>
<li> Assorted new test cases and test case fixes (endian and word
size issues).</li>
<li> Fixes for big-endian codegen bugs, instruction encodings, and
instruction constraints.</li>
<li> Implemented -integrated-as support.</li>
<li> Additional support for Altivec compare operations.</li>
<li> IBM long double support.</li>
</ul>
<p>There have also been code generation improvements for both 32- and 64-bit
code. Instruction scheduling support for the Freescale e500mc and e5500
cores has been added.</p>
</div>
<!--=========================================================================-->
<h3>
<a name="NVPTX">PTX/NVPTX Target Improvements</a>
</h3>
<div>
<p>The PTX back-end has been replaced by the NVPTX back-end, which is based on
the LLVM back-end used by NVIDIA in their CUDA (nvcc) and OpenCL compiler.
Some highlights include:</p>
<ul>
<li>Compatibility with PTX 3.1 and SM 3.5</li>
<li>Support for NVVM intrinsics as defined in the NVIDIA Compiler SDK</li>
<li>Full compatibility with old PTX back-end, with much greater coverage of
LLVM IR</li>
</ul>
<p>Please submit any back-end bugs to the LLVM Bugzilla site.</p>
</div>
<!--=========================================================================-->
<h3>
<a name="OtherTS">Other Target Specific Improvements</a>
</h3>
<div>
<ul>
<li>Added support for custom names for library functions in TargetLibraryInfo.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="changes">Major Changes and Removed Features</a>
</h3>
<div>
<p>If you're already an LLVM user or developer with out-of-tree changes based on
LLVM 3.2, this section lists some "gotchas" that you may run into upgrading
from the previous release.</p>
<ul>
<li>llvm-ld and llvm-stub have been removed, llvm-ld functionality can be partially replaced by
llvm-link | opt | {llc | as, llc -filetype=obj} | ld, or fully replaced by Clang. </li>
<li>MCJIT: added support for inline assembly (requires asm parser), added faux remote target execution to lli option '-remote-mcjit'.</li>
</ul>
</div>
<!--=========================================================================-->
<h3>
<a name="api_changes">Internal API Changes</a>
</h3>
<div>
<p>In addition, many APIs have changed in this release. Some of the major
LLVM API changes are:</p>
<p> We've added a new interface for allowing IR-level passes to access
target-specific information. A new IR-level pass, called
"TargetTransformInfo" provides a number of low-level interfaces.
LSR and LowerInvoke already use the new interface. </p>
<p> The TargetData structure has been renamed to DataLayout and moved to VMCore
to remove a dependency on Target. </p>
</div>
<!--=========================================================================-->
<h3>
<a name="tools_changes">Tools Changes</a>
</h3>
<div>
<p>In addition, some tools have changed in this release. Some of the changes are:</p>
<ul>
<li>opt: added support for '-mtriple' option.</li>
<li>llvm-mc : - added '-disassemble' support for '-show-inst' and '-show-encoding' options, added '-edis' option to produce annotated
disassembly output for X86 and ARM targets.</li>
<li>libprofile: allows the profile data file name to be specified by the LLVMPROF_OUTPUT environment variable.</li>
<li>llvm-objdump: has been changed to display available targets, '-arch' option accepts x86 and x86-64 as valid arch names.</li>
<li>llc and opt: added FMA formation from pairs of FADD + FMUL or FSUB + FMUL enabled by option '-enable-excess-fp-precision' or option '-enable-unsafe-fp-math',
option '-fp-contract' controls the creation by optimizations of fused FP by selecting Fast, Standard, or Strict mode.</li>
<li>llc: object file output from llc is no longer considered experimental.</li>
<li>gold plugin: handles Position Independent Executables.</li>
</ul>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="knownproblems">Known Problems</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>LLVM is generally a production quality compiler, and is used by a broad range
of applications and shipping in many products. That said, not every
subsystem is as mature as the aggregate, particularly the more obscure
targets. If you run into a problem, please check
the <a href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
there isn't already one or ask on
the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev
list</a>.</p>
<p>Known problem areas include:</p>
<ul>
<li>The CellSPU, MSP430, and XCore backends are experimental, and the CellSPU backend will be removed in LLVM 3.3.</li>
<li>The integrated assembler, disassembler, and JIT is not supported by
several targets. If an integrated assembler is not supported, then a
system assembler is required. For more details, see the <a
href="CodeGenerator.html#targetfeatures">Target Features Matrix</a>.
</li>
</ul>
</div>
<!-- *********************************************************************** -->
<h2>
<a name="additionalinfo">Additional Information</a>
</h2>
<!-- *********************************************************************** -->
<div>
<p>A wide variety of additional information is available on
the <a href="http://llvm.org/">LLVM web page</a>, in particular in
the <a href="http://llvm.org/docs/">documentation</a> section. The web page
also contains versions of the API documentation which is up-to-date with the
Subversion version of the source code. You can access versions of these
documents specific to this release by going into the "<tt>llvm/doc/</tt>"
directory in the LLVM tree.</p>
<p>If you have any questions or comments about LLVM, please feel free to contact
us via the <a href="http://llvm.org/docs/#maillist"> mailing lists</a>.</p>
</div>
<!-- *********************************************************************** -->
<hr>
<address>
<a href="http://jigsaw.w3.org/css-validator/check/referer"><img
src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
<a href="http://validator.w3.org/check/referer"><img
src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
<a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
Last modified: $Date$
</address>
</body>
</html>