1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
6 <link rel="stylesheet" href="_static/llvm.css" type="text/css">
7 <title>LLVM 3.1 Release Notes</title>
11 <h1>LLVM 3.1 Release Notes</h1>
14 <img style="float:right" src="http://llvm.org/img/DragonSmall.png"
15 width="136" height="136" alt="LLVM Dragon Logo">
19 <li><a href="#intro">Introduction</a></li>
20 <li><a href="#subproj">Sub-project Status Update</a></li>
21 <li><a href="#externalproj">External Projects Using LLVM 3.1</a></li>
22 <li><a href="#whatsnew">What's New in LLVM?</a></li>
23 <li><a href="GettingStarted.html">Installation Instructions</a></li>
24 <li><a href="#knownproblems">Known Problems</a></li>
25 <li><a href="#additionalinfo">Additional Information</a></li>
28 <div class="doc_author">
29 <p>Written by the <a href="http://llvm.org/">LLVM Team</a></p>
32 <h1 style="color:red">These are in-progress notes for the upcoming LLVM 3.1
35 <a href="http://llvm.org/releases/3.0/docs/ReleaseNotes.html">LLVM 3.0
36 Release Notes</a>.</h1>
38 <!-- *********************************************************************** -->
40 <a name="intro">Introduction</a>
42 <!-- *********************************************************************** -->
46 <p>This document contains the release notes for the LLVM Compiler
47 Infrastructure, release 3.1. Here we describe the status of LLVM, including
48 major improvements from the previous release, improvements in various
49 subprojects of LLVM, and some of the current users of the code.
50 All LLVM releases may be downloaded from
51 the <a href="http://llvm.org/releases/">LLVM releases web site</a>.</p>
53 <p>For more information about LLVM, including information about the latest
54 release, please check out the <a href="http://llvm.org/">main LLVM web
55 site</a>. If you have questions or comments,
56 the <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM
57 Developer's Mailing List</a> is a good place to send them.</p>
59 <p>Note that if you are reading this file from a Subversion checkout or the main
60 LLVM web page, this document applies to the <i>next</i> release, not the
61 current one. To see the release notes for a specific release, please see the
62 <a href="http://llvm.org/releases/">releases page</a>.</p>
67 <!-- *********************************************************************** -->
69 <a name="subproj">Sub-project Status Update</a>
71 <!-- *********************************************************************** -->
75 <p>The LLVM 3.1 distribution currently consists of code from the core LLVM
76 repository (which roughly includes the LLVM optimizers, code generators and
77 supporting tools), and the Clang repository. In addition to this code, the
78 LLVM Project includes other sub-projects that are in development. Here we
79 include updates on these subprojects.</p>
81 <!--=========================================================================-->
83 <a name="clang">Clang: C/C++/Objective-C Frontend Toolkit</a>
88 <p><a href="http://clang.llvm.org/">Clang</a> is an LLVM front end for the C,
89 C++, and Objective-C languages. Clang aims to provide a better user
90 experience through expressive diagnostics, a high level of conformance to
91 language standards, fast compilation, and low memory use. Like LLVM, Clang
92 provides a modular, library-based architecture that makes it suitable for
93 creating or integrating with other development tools. Clang is considered a
94 production-quality compiler for C, Objective-C, C++ and Objective-C++ on x86
95 (32- and 64-bit), and for Darwin/ARM targets.</p>
97 <p>In the LLVM 3.1 time-frame, the Clang team has made many improvements.
98 Highlights include:</p>
100 <li>Greatly expanded <a href="http://clang.llvm.org/cxx_status.html">C++11
101 support</a> including lambdas, initializer lists, constexpr, user-defined
102 literals, and atomics.</li>
103 <li>A new <a href="http://clang.llvm.org/docs/Tooling.html">tooling</a>
104 library to ease building of clang-based standalone tools.</li>
105 <li>Extended support for
106 <a href="http://clang.llvm.org/docs/ObjectiveCLiterals.html">literals in
107 Objective C</a>.</li>
110 <p>For more details about the changes to Clang since the 3.0 release, see the
111 <a href="http://clang.llvm.org/docs/ReleaseNotes.html">Clang release
114 <p>If Clang rejects your code but another compiler accepts it, please take a
115 look at the <a href="http://clang.llvm.org/compatibility.html">language
116 compatibility</a> guide to make sure this is not intentional or a known
121 <!--=========================================================================-->
123 <a name="dragonegg">DragonEgg: GCC front-ends, LLVM back-end</a>
128 <p><a href="http://dragonegg.llvm.org/">DragonEgg</a> is a
129 <a href="http://gcc.gnu.org/wiki/plugins">gcc plugin</a> that replaces GCC's
130 optimizers and code generators with LLVM's. It works with gcc-4.5 and gcc-4.6
131 (and partially with gcc-4.7), can target the x86-32/x86-64 and ARM processor
132 families, and has been successfully used on the Darwin, FreeBSD, KFreeBSD,
133 Linux and OpenBSD platforms. It fully supports Ada, C, C++ and Fortran. It
134 has partial support for Go, Java, Obj-C and Obj-C++.</p>
136 <p>The 3.1 release has the following notable changes:</p>
139 <li>Partial support for gcc-4.7. Ada support is poor, but other languages work
142 <li>Support for ARM processors. Some essential gcc headers that are needed to
143 build DragonEgg for ARM are not installed by gcc. To work around this,
144 copy the missing headers from the gcc source tree.</li>
146 <li>Better optimization for Fortran by exploiting the fact that Fortran scalar
147 arguments have 'restrict' semantics.</li>
149 <li>Better optimization for all languages by passing information about type
150 aliasing and type ranges to the LLVM optimizers.</li>
152 <li>A regression test-suite was added.</li>
157 <!--=========================================================================-->
159 <a name="compiler-rt">compiler-rt: Compiler Runtime Library</a>
164 <p>The new LLVM <a href="http://compiler-rt.llvm.org/">compiler-rt project</a>
165 is a simple library that provides an implementation of the low-level
166 target-specific hooks required by code generation and other runtime
167 components. For example, when compiling for a 32-bit target, converting a
168 double to a 64-bit unsigned integer is compiled into a runtime call to the
169 "__fixunsdfdi" function. The compiler-rt library provides highly optimized
170 implementations of this and other low-level routines (some are 3x faster than
171 the equivalent libgcc routines).</p>
175 <!--=========================================================================-->
177 <a name="lldb">LLDB: Low Level Debugger</a>
182 <p><a href="http://lldb.llvm.org">LLDB</a> is a ground-up implementation of a
183 command line debugger, as well as a debugger API that can be used from other
184 applications. LLDB makes use of the Clang parser to provide high-fidelity
185 expression parsing (particularly for C++) and uses the LLVM JIT for target
190 <!--=========================================================================-->
192 <a name="libc++">libc++: C++ Standard Library</a>
197 <p>Like compiler_rt, libc++ is now <a href="DeveloperPolicy.html#license">dual
198 licensed</a> under the MIT and UIUC license, allowing it to be used more
203 <!--=========================================================================-->
205 <a name="vmkit">VMKit</a>
210 <p>The <a href="http://vmkit.llvm.org/">VMKit project</a> is an implementation
211 of a Java Virtual Machine (Java VM or JVM) that uses LLVM for static and
212 just-in-time compilation.</p>
214 <p>In the LLVM 3.1 time-frame, VMKit has had significant improvements on both
215 runtime and startup performance.</p>
220 <!--=========================================================================-->
222 <a name="Polly">Polly: Polyhedral Optimizer</a>
227 <p><a href="http://polly.llvm.org/">Polly</a> is an <em>experimental</em>
228 optimizer for data locality and parallelism. It currently provides high-level
229 loop optimizations and automatic parallelisation (using the OpenMP run time).
230 Work in the area of automatic SIMD and accelerator code generation was
233 <p>Within the LLVM 3.1 time-frame there were the following highlights:</p>
236 <li>Polly became an official LLVM project</li>
237 <li>Polly can be loaded directly into clang (enabled by '-O3 -mllvm -polly')</li>
238 <li>An automatic scheduling optimizer (derived
239 from <a href="http://pluto-compiler.sourceforge.net/">Pluto</a>) was
240 integrated. It performs loop transformations to optimize for data-locality
241 and parallelism. The transformations include, but are not limited to
242 interchange, fusion, fission, skewing and tiling.</li>
249 <!-- *********************************************************************** -->
251 <a name="externalproj">External Open Source Projects Using LLVM 3.1</a>
253 <!-- *********************************************************************** -->
257 <p>An exciting aspect of LLVM is that it is used as an enabling technology for
258 a lot of other language and tools projects. This section lists some of the
259 projects that have already been updated to work with LLVM 3.1.</p>
265 <p><a href="http://faust.grame.fr/">FAUST</a> is a compiled language for
266 real-time audio signal processing. The name FAUST stands for Functional
267 AUdio STream. Its programming model combines two approaches: functional
268 programming and block diagram composition. In addition with the C, C++, Java,
269 JavaScript output formats, the Faust compiler can generate LLVM bitcode, and
270 works with LLVM 2.7-3.1.</p>
274 <h3>Glasgow Haskell Compiler (GHC)</h3>
278 <p><a href="http://www.haskell.org/ghc/">GHC</a> is an open source compiler and
279 programming suite for Haskell, a lazy functional programming language. It
280 includes an optimizing static compiler generating good code for a variety of
281 platforms, together with an interactive system for convenient, quick
284 <p>GHC 7.0 and onwards include an LLVM code generator, supporting LLVM 2.8 and
293 <p><a href="https://github.com/JuliaLang/julia">Julia</a> is a high-level,
294 high-performance dynamic language for technical computing. It provides a
295 sophisticated compiler, distributed parallel execution, numerical accuracy,
296 and an extensive mathematical function library. The compiler uses type
297 inference to generate fast code without any type declarations, and uses
298 LLVM's optimization passes and JIT compiler. The
299 <a href="http://julialang.org/"> Julia Language</a> is designed
300 around multiple dispatch, giving programs a large degree of flexibility. It
301 is ready for use on many kinds of problems.</p>
305 <h3>LLVM D Compiler</h3>
309 <p><a href="https://github.com/ldc-developers/ldc">LLVM D Compiler</a> (LDC) is
310 a compiler for the D programming Language. It is based on the DMD frontend
311 and uses LLVM as backend.</p>
315 <h3>Open Shading Language</h3>
319 <p><a href="https://github.com/imageworks/OpenShadingLanguage/">Open Shading
320 Language (OSL)</a> is a small but rich language for programmable shading in
321 advanced global illumination renderers and other applications, ideal for
322 describing materials, lights, displacement, and pattern generation. It uses
323 LLVM to JIT complex shader networks to x86 code at runtime.</p>
325 <p>OSL was developed by Sony Pictures Imageworks for use in its in-house
326 renderer used for feature film animation and visual effects, and is
327 distributed as open source software with the "New BSD" license.</p>
331 <h3>Portable OpenCL (pocl)</h3>
335 <p>In addition to producing an easily portable open source OpenCL
336 implementation, another major goal of <a href="http://pocl.sourceforge.net/">
337 pocl</a> is improving performance portability of OpenCL programs with
338 compiler optimizations, reducing the need for target-dependent manual
339 optimizations. An important part of pocl is a set of LLVM passes used to
340 statically parallelize multiple work-items with the kernel compiler, even in
341 the presence of work-group barriers. This enables static parallelization of
342 the fine-grained static concurrency in the work groups in multiple ways
343 (SIMD, VLIW, superscalar,...).</p>
351 <p><a href="http://pure-lang.googlecode.com/">Pure</a> is an
352 algebraic/functional programming language based on term rewriting. Programs
353 are collections of equations which are used to evaluate expressions in a
354 symbolic fashion. The interpreter uses LLVM as a backend to JIT-compile Pure
355 programs to fast native code. Pure offers dynamic typing, eager and lazy
356 evaluation, lexical closures, a hygienic macro system (also based on term
357 rewriting), built-in list and matrix support (including list and matrix
358 comprehensions) and an easy-to-use interface to C and other programming
359 languages (including the ability to load LLVM bitcode modules, and inline C,
360 C++, Fortran and Faust code in Pure programs if the corresponding
361 LLVM-enabled compilers are installed).</p>
363 <p>Pure version 0.54 has been tested and is known to work with LLVM 3.1 (and
364 continues to work with older LLVM releases >= 2.5).</p>
368 <h3>TTA-based Co-design Environment (TCE)</h3>
372 <p><a href="http://tce.cs.tut.fi/">TCE</a> is a toolset for designing
373 application-specific processors (ASP) based on the Transport triggered
374 architecture (TTA). The toolset provides a complete co-design flow from C/C++
375 programs down to synthesizable VHDL/Verilog and parallel program binaries.
376 Processor customization points include the register files, function units,
377 supported operations, and the interconnection network.</p>
379 <p>TCE uses Clang and LLVM for C/C++ language support, target independent
380 optimizations and also for parts of code generation. It generates new
381 LLVM-based code generators "on the fly" for the designed TTA processors and
382 loads them in to the compiler backend as runtime libraries to avoid
383 per-target recompilation of larger parts of the compiler chain.</p>
389 <!-- *********************************************************************** -->
391 <a name="whatsnew">What's New in LLVM 3.1?</a>
393 <!-- *********************************************************************** -->
397 <p>This release includes a huge number of bug fixes, performance tweaks and
398 minor improvements. Some of the major improvements and new features are
399 listed in this section.</p>
401 <!--=========================================================================-->
403 <a name="majorfeatures">Major New Features</a>
408 <!-- Features that need text if they're finished for 3.1:
412 loop dependence analysis
413 CorrelatedValuePropagation
414 lib/Transforms/IPO/MergeFunctions.cpp => consider for 3.1.
415 Integrated assembler on by default for arm/thumb?
420 Analysis/RegionInfo.h + Dom Frontiers
421 SparseBitVector: used in LiveVar.
422 llvm/lib/Archive - replace with lib object?
425 <p>LLVM 3.1 includes several major changes and big features:</p>
428 <li><a href="../tools/clang/docs/AddressSanitizer.html">AddressSanitizer</a>,
429 a fast memory error detector.</li>
430 <li><a href="CodeGenerator.html#machineinstrbundle">MachineInstr Bundles</a>,
431 Support to model instruction bundling / packing.</li>
432 <li><a href="#armintegratedassembler">ARM Integrated Assembler</a>,
433 A full featured assembler and direct-to-object support for ARM.</li>
434 <li><a href="#blockplacement">Basic Block Placement</a>
435 Probability driven basic block placement.</li>
441 <!--=========================================================================-->
443 <a name="coreimprovements">LLVM IR and Core Improvements</a>
448 <p>LLVM IR has several new features for better support of new targets and that
449 expose new optimization opportunities:</p>
452 <li>A new type representing 16 bit <i>half</i> floating point values has
454 <li>IR now supports vectors of pointers, including vector GEPs.</li>
455 <li>Module flags have been introduced. They convey information about the
456 module as a whole to LLVM subsystems. This is currently used to encode
457 Objective C ABI information.</li>
458 <li>Loads can now have range metadata attached to them to describe the
459 possible values being loaded.</li>
460 <li>The <tt>llvm.ctlz</tt> and <tt>llvm.cttz</tt> intrinsics now have an
461 additional argument which indicates whether the behavior of the intrinsic
462 is undefined on a zero input. This can be used to generate more efficient
463 code on platforms that only have instructions which don't return the type
464 size when counting bits in 0.</li>
469 <!--=========================================================================-->
471 <a name="optimizer">Optimizer Improvements</a>
476 <p>In addition to many minor performance tweaks and bug fixes, this
477 release includes a few major enhancements and additions to the
481 <li>The loop unroll pass now is able to unroll loops with run-time trip counts.
482 This feature is turned off by default, and is enabled with the
483 <code>-unroll-runtime</code> flag.</li>
484 <li>A new basic-block autovectorization pass is available. Pass
485 <code>-vectorize</code> to run this pass along with some associated
486 post-vectorization cleanup passes. For more information, see the EuroLLVM
487 2012 slides: <a href="http://llvm.org/devmtg/2012-04-12/Slides/Hal_Finkel.pdf">
488 Autovectorization with LLVM</a>.</li>
489 <li>Inline cost heuristics have been completely overhauled and now closely
490 model constant propagation through call sites, disregard trivially dead
491 code costs, and can model C++ STL iterator patterns.</li>
496 <!--=========================================================================-->
498 <a name="mc">MC Level Improvements</a>
503 <p>The LLVM Machine Code (aka MC) subsystem was created to solve a number of
504 problems in the realm of assembly, disassembly, object file format handling,
505 and a number of other related areas that CPU instruction-set level tools work
506 in. For more information, please see
507 the <a href="http://blog.llvm.org/2010/04/intro-to-llvm-mc-project.html">Intro
508 to the LLVM MC Project Blog Post</a>.</p>
511 <li>The integrated assembler can optionally emit debug information when
512 assembling a </tt>.s</tt> file. It can be enabled by passing the
513 <tt>-g</tt> option to <tt>llvm-mc</tt>.</li>
518 <!--=========================================================================-->
520 <a name="codegen">Target Independent Code Generator Improvements</a>
525 <p>We have changed the way that the Type Legalizer legalizes vectors. The type
526 legalizer now attempts to promote integer elements. This enabled the
527 implementation of vector-select. Additionally, we see a performance boost on
528 workloads which use vectors of chars and shorts, since they are now promoted
529 to 32-bit types, which are better supported by the SIMD instruction set.
530 Floating point types are still widened as before.</p>
533 <p>We have put a significant amount of work into the code generator
534 infrastructure, which allows us to implement more aggressive algorithms and
535 make it run faster:</p>
538 <li>TableGen can now synthesize register classes that are only needed to
539 represent combinations of constraints from instructions and sub-registers.
540 The synthetic register classes inherit most of their properties form their
541 closest user-defined super-class.</li>
542 <li><code>MachineRegisterInfo</code> now allows the reserved registers to be
543 frozen when register allocation starts. Target hooks should use the
544 <code>MRI->canReserveReg(FramePtr)</code> method to avoid accidentally
545 disabling frame pointer elimination during register allocation.</li>
546 <li>A new kind of <code>MachineOperand</code> provides a compact
547 representation of large clobber lists on call instructions. The register
548 mask operand references a bit mask of preserved registers. Everything else
550 <li>The DWARF debug info writer gained support for emitting data for the
551 <a href="SourceLevelDebugging.html#acceltable">name accelerator tables
552 DWARF extension</a>. It is used by LLDB to speed up name lookup.</li>
555 <p> We added new TableGen infrastructure to support bundling for
556 Very Long Instruction Word (VLIW) architectures. TableGen can now
557 automatically generate a deterministic finite automaton from a VLIW
558 target's schedule description which can be queried to determine
559 legal groupings of instructions in a bundle.</p>
561 <p> We have added a new target independent VLIW packetizer based on the
562 DFA infrastructure to group machine instructions into bundles.</p>
567 <a name="blockplacement">Basic Block Placement</a>
570 <p>A probability based block placement and code layout algorithm was added to
571 LLVM's code generator. This layout pass supports probabilities derived from
572 static heuristics as well as source code annotations such as
573 <code>__builtin_expect</code>.</p>
576 <!--=========================================================================-->
578 <a name="x86">X86-32 and X86-64 Target Improvements</a>
583 <p>New features and major changes in the X86 target include:</p>
586 <li>Greatly improved support for AVX2.</li>
587 <li>Lots of bug fixes and improvements for AVX1.</li>
588 <li>Support for the FMA4 and XOP instruction set extensions.</li>
589 <li>Call instructions use the new register mask operands for faster compile
590 times and better support for different calling conventions. The old WINCALL
591 instructions are no longer needed.</li>
592 <li>DW2 Exception Handling is enabled on Cygwin and MinGW.</li>
593 <li>Support for implicit TLS model used with MSVC runtime.</li>
598 <!--=========================================================================-->
600 <a name="ARM">ARM Target Improvements</a>
605 <p>New features of the ARM target include:</p>
608 <li>The constant island pass now supports basic block and constant pool entry
609 alignments greater than 4 bytes.</li>
610 <li>On Darwin, the ARM target now has a full-featured integrated assembler.
615 <a name="armintegratedassembler">ARM Integrated Assembler</a>
618 <p>The ARM target now includes a full featured macro assembler, including
619 direct-to-object module support for clang. The assembler is currently enabled
620 by default for Darwin only pending testing and any additional necessary
621 platform specific support for Linux.</p>
623 <p>Full support is included for Thumb1, Thumb2 and ARM modes, along with
624 subtarget and CPU specific extensions for VFP2, VFP3 and NEON.</p>
626 <p>The assembler is Unified Syntax only (see ARM Architecural Reference Manual
627 for details). While there is some, and growing, support for pre-unfied (divided)
628 syntax, there are still significant gaps in that support.</p>
632 <!--=========================================================================-->
634 <a name="MIPS">MIPS Target Improvements</a>
638 New features and major changes in the MIPS target include:</p>
641 <li>MIPS32 little-endian direct object code emission is functional.</li>
642 <li>MIPS64 little-endian code generation is largely functional for N64 ABI in assembly printing mode with the exception of handling of long double (f128) type.</li>
643 <li>Support for new instructions has been added, which includes swap-bytes
644 instructions (WSBH and DSBH), floating point multiply-add/subtract and
645 negative multiply-add/subtract instructions, and floating
646 point load/store instructions with reg+reg addressing (LWXC1, etc.)</li>
647 <li>Various fixes to improve performance have been implemented.</li>
648 <li>Post-RA scheduling is now enabled at -O3.</li>
649 <li>Support for soft-float code generation has been added.</li>
653 <!--=========================================================================-->
655 <a name="PTX">PTX Target Improvements</a>
660 <p>An outstanding conditional inversion bug was fixed in this release.</p>
662 <p><b>NOTE</b>: LLVM 3.1 marks the last release of the PTX back-end, in its
663 current form. The back-end is currently being replaced by the NVPTX
664 back-end, currently in SVN ToT.</p>
668 <!--=========================================================================-->
670 <a name="OtherTS">Other Target Specific Improvements</a>
676 <li>Support for Qualcomm's Hexagon VLIW processor has been added.</li>
681 <!--=========================================================================-->
683 <a name="changes">Major Changes and Removed Features</a>
688 <p>If you're already an LLVM user or developer with out-of-tree changes based on
689 LLVM 3.1, this section lists some "gotchas" that you may run into upgrading
690 from the previous release.</p>
693 <li>LLVM's build system now requires a python 2 interpreter to be present at
694 build time. A perl interpreter is no longer required.</li>
695 <li>The C backend has been removed. It had numerous problems, to the point of
696 not being able to compile any nontrivial program.</li>
697 <li>The Alpha, Blackfin and SystemZ targets have been removed due to lack of
699 <li>LLVM 3.1 removes support for reading LLVM 2.9 bitcode files. Going
700 forward, we aim for all future versions of LLVM to read bitcode files and
701 <tt>.ll</tt> files produced by LLVM 3.0 and later.</li>
702 <li>The <tt>unwind</tt> instruction is now gone. With the introduction of the
703 new exception handling system in LLVM 3.0, the <tt>unwind</tt> instruction
704 became obsolete.</li>
705 <li>LLVM 3.0 and earlier automatically added the returns_twice fo functions
706 like setjmp based on the name. This functionality was removed in 3.1.
707 This affects Clang users, if -ffreestanding is used.</li>
712 <!--=========================================================================-->
714 <a name="api_changes">Internal API Changes</a>
719 <p>In addition, many APIs have changed in this release. Some of the major
720 LLVM API changes are:</p>
723 <li>Target specific options have been moved from global variables to members
724 on the new <code>TargetOptions</code> class, which is local to each
725 <code>TargetMachine</code>. As a consequence, the associated flags will
726 no longer be accepted by <tt>clang -mllvm</tt>. This includes:
728 <li><code>llvm::PrintMachineCode</code></li>
729 <li><code>llvm::NoFramePointerElim</code></li>
730 <li><code>llvm::NoFramePointerElimNonLeaf</code></li>
731 <li><code>llvm::DisableFramePointerElim(const MachineFunction &)</code></li>
732 <li><code>llvm::LessPreciseFPMADOption</code></li>
733 <li><code>llvm::LessPrecideFPMAD()</code></li>
734 <li><code>llvm::NoExcessFPPrecision</code></li>
735 <li><code>llvm::UnsafeFPMath</code></li>
736 <li><code>llvm::NoInfsFPMath</code></li>
737 <li><code>llvm::NoNaNsFPMath</code></li>
738 <li><code>llvm::HonorSignDependentRoundingFPMathOption</code></li>
739 <li><code>llvm::HonorSignDependentRoundingFPMath()</code></li>
740 <li><code>llvm::UseSoftFloat</code></li>
741 <li><code>llvm::FloatABIType</code></li>
742 <li><code>llvm::NoZerosInBSS</code></li>
743 <li><code>llvm::JITExceptionHandling</code></li>
744 <li><code>llvm::JITEmitDebugInfo</code></li>
745 <li><code>llvm::JITEmitDebugInfoToDisk</code></li>
746 <li><code>llvm::GuaranteedTailCallOpt</code></li>
747 <li><code>llvm::StackAlignmentOverride</code></li>
748 <li><code>llvm::RealignStack</code></li>
749 <li><code>llvm::DisableJumpTables</code></li>
750 <li><code>llvm::EnableFastISel</code></li>
751 <li><code>llvm::getTrapFunctionName()</code></li>
752 <li><code>llvm::EnableSegmentedStacks</code></li>
755 <li>The <code>MDBuilder</code> class has been added to simplify the creation
761 <!--=========================================================================-->
763 <a name="tools_changes">Tools Changes</a>
768 <p>In addition, some tools have changed in this release. Some of the changes
773 <li><tt>llvm-stress</tt> is a command line tool for generating random
774 <tt>.ll</tt> files to fuzz different LLVM components. </li>
775 <li>The <tt>llvm-ld</tt> tool has been removed. The clang driver provides a
776 more reliable solution for turning a set of bitcode files into a binary.
777 To merge bitcode files <tt>llvm-link</tt> can be used instead.</li>
783 <!--=========================================================================-->
785 <a name="python">Python Bindings</a>
790 <p>Officially supported Python bindings have been added! Feature support is far
791 from complete. The current bindings support interfaces to:</p>
793 <li>Object File Interface</li>
794 <li>Disassembler</li>
797 <p>Using the Object File Interface, it is possible to inspect binary object files.
798 Think of it as a Python version of readelf or llvm-objdump.</p>
800 <p>Support for additional features is currently being developed by community
801 contributors. If you are interested in shaping the direction of the Python
802 bindings, please express your intent on IRC or the developers list.</p>
808 <!-- *********************************************************************** -->
810 <a name="knownproblems">Known Problems</a>
812 <!-- *********************************************************************** -->
816 <p>LLVM is generally a production quality compiler, and is used by a broad range
817 of applications and shipping in many products. That said, not every
818 subsystem is as mature as the aggregate, particularly the more obscure
819 targets. If you run into a problem, please check the <a
820 href="http://llvm.org/bugs/">LLVM bug database</a> and submit a bug if
821 there isn't already one or ask on the <a
822 href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVMdev
825 <p>Known problem areas include:</p>
828 <li>The CellSPU, MSP430, PTX and XCore backends are experimental.</li>
830 <li>The integrated assembler, disassembler, and JIT is not supported by
831 several targets. If an integrated assembler is not supported, then a
832 system assembler is required. For more details, see the <a
833 href="CodeGenerator.html#targetfeatures">Target Features Matrix</a>.
839 <!-- *********************************************************************** -->
841 <a name="additionalinfo">Additional Information</a>
843 <!-- *********************************************************************** -->
847 <p>A wide variety of additional information is available on
848 the <a href="http://llvm.org/">LLVM web page</a>, in particular in
849 the <a href="http://llvm.org/docs/">documentation</a> section. The web page
850 also contains versions of the API documentation which is up-to-date with the
851 Subversion version of the source code. You can access versions of these
852 documents specific to this release by going into the "<tt>llvm/doc/</tt>"
853 directory in the LLVM tree.</p>
855 <p>If you have any questions or comments about LLVM, please feel free to contact
856 us via the <a href="http://llvm.org/docs/#maillist"> mailing lists</a>.</p>
860 <!-- *********************************************************************** -->
864 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
865 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
866 <a href="http://validator.w3.org/check/referer"><img
867 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
869 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
870 Last modified: $Date$