diff options
| -rw-r--r-- | docs/ReleaseNotes.html | 154 |
1 files changed, 88 insertions, 66 deletions
diff --git a/docs/ReleaseNotes.html b/docs/ReleaseNotes.html index c8c0bd375f..ebc0fd7374 100644 --- a/docs/ReleaseNotes.html +++ b/docs/ReleaseNotes.html @@ -66,6 +66,13 @@ It includes many features and refinements from LLVM 2.1.</p> </div> +<!-- Unfinished features in 2.2: + Index Set Splitting not enabled by default + Machine LICM + Machine Sinking + LegalizeDAGTypes + --> + <!--=========================================================================--> <div class="doc_subsection"> <a name="deprecation">Deprecated features in LLVM 2.2</a> @@ -92,10 +99,10 @@ llvm 2.2, and will be redesigned or removed in llvm 2.3.</p> LLVM 2.1, llvm-gcc 4.2 was beta). Since LLVM 2.1, the llvm-gcc 4.2 front-end has made leaps and bounds and is now at least as good as 4.0 in virtually every area, and is better in several areas (for example, exception handling -correctness). We strongly recommend that you migrate from llvm-gcc 4.0 to -llvm-gcc 4.2 in this release cycle because <b>LLVM 2.2 is the last release -that will support llvm-gcc 4.0</b>: LLVM 2.3 will only support the llvm-gcc -4.2 front-end.</p> +correctness, support for Ada and FORTRAN). We strongly recommend that you +migrate from llvm-gcc 4.0 to llvm-gcc 4.2 in this release cycle because +<b>LLVM 2.2 is the last release that will support llvm-gcc 4.0</b>: LLVM 2.3 +will only support the llvm-gcc 4.2 front-end.</p> <p>The <a href="http://clang.llvm.org/">clang project</a> is an effort to build a set of new 'llvm native' front-end technologies for the LLVM optimizer @@ -104,7 +111,7 @@ nicely, and it has advanced source-to-source analysis and transformation capabilities. If you are interested in building source-level tools for C and Objective-C (and eventually C++), you should take a look. However, note that clang is not an official part of the LLVM 2.2 release. If you are interested in -this project, please see the web site.</p> +this project, please see its <a href="http://clang.llvm.org/">web site</a>.</p> </div> @@ -115,88 +122,113 @@ this project, please see the web site.</p> <div class="doc_text"> -<p>Scott Michel contributed Cell backend</p> - -<p>Dale contributed full support for long double on x86/x86-64 (where it is 80 -bits) and on Darwin PPC/PPC64 (where it is 128 bits).</p> +<p>LLVM 2.2 includes several major new capabilities:</p> -<p>Ada, gfortran</p> +<ul> +<li>Scott Michel contributed an SPU backend, which generates code for the +vector coprocessors on the Cell processor. (Status?)</li> -<p> -debug improvements -O0 +<li>llvm-gcc 4.2 has significantly improved support for the GCC Ada (GNAT) and +FORTRAN (gfortran) frontends. Duncan has the llvm-gcc 4.2 GNAT front-end +supporting almost all of the ACATS testsuite (except 2 tests?). The llvm-gcc +4.2 gfortran front-end supports a broad range of FORTRAN code, but does <a +href="http://llvm.org/PR1971">not support EQUIVALENCE yet</a>.</li> -Gordon: GC Revamp. docs/GarbageCollection.html +<li>Dale contributed full support for long double on x86/x86-64 (where it is 80 +bits) and on Darwin PPC/PPC64 (where it is 128 bits). In previous LLVM +releases, llvm-gcc silently mapped long double to double.</li> -Kaleidoscope: docs/tutorial +<li>Gordon rewrote most of the <a href="GarbageCollection.html">Accurate Garbage +Collection</a> code in the code generator, making the generated code more +efficient and adding support for the Ocaml garbage collector metadata +format.</li> -Gordon: C and Ocaml Bindings +<li>LLVM now includes a new set of detailed <a +href="tutorial/index.html">tutorials</a>, which explain how to implement a +language with LLVM and shows how to use several important APIs.</li> -Christopher Lamb: Multiple address spaces. </div> <!--=========================================================================--> <div class="doc_subsection"> -<a name="optimizer">Optimizer Improvements</a> +<a name="codegen">Code Generator Improvements</a> </div> <div class="doc_text"> -<p>Some of the most noticable feature improvements this release have been in the -optimizer, speeding it up and making it more aggressive. For example:</p> +<p>We 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> <ul> -<li>Daniel Berlin and Curtis Dunham rewrote Andersen's alias analysis (which is -not enabled by default) to be several orders of magnitude faster, implemented -Offline Variable Substitution and Lazy Cycle Detection.</li> +<li>Owen refactored the existing LLVM dominator and loop information code to +allow it work on the machine code representation. He contributed support for +dominator and loop information on machine code and merged the code for forward +and backward dominator computation.</li> -Devang: LoopIndexSplit is enabled by default. +<li>Dan added support for emitting debug information with .file and .loc +directives on that support it, instead of emitting large tables in the .s +file.</li> -Dan Gohman: LSR improvements for SSE intrinsics and +<li>Evan extended the DAG scheduler to model physical register dependencies +explicitly and have the BURR scheduler pick a correct schedule based on the +dependencies. This reduces our use of the 'flag' operand hack.</li> -Evan added simple exit value substitution to LSR. +<li>Evan added initial support for register coalescing of subregister +references.</li> -Anton added readnone/readonly attributes for modeling function side effects -and Duncan hooked up GCC's pure/const attributes to use them and enhanced alias -analysis to use them. +<li>Rafael Espindola implemented initial support for a new 'byval' attribute, +which allows more efficient by-value argument passing in the LLVM IR. Evan +finished support for it and enabled it in the X86 (32- and 64-bit) and C +backends.</li> + +<li>The LLVM TargetInstrInfo class can now answer queries about the mod/ref and +side-effect behavior of MachineInstr's. This information is inferred +automatically by TableGen from .td files for all instructions with +patterns.</li> + +<li>Evan implemented simple live interval splitting on basic block boundaries. +This allows the register allocator to be more successful at keeping values in +registers in some parts of a value's live range, even if they need to be spilled +in some other block.</li> + +<li>The new MachineRegisterInfo.h class provides support for efficiently +iterating over all defs/uses of a register, and this information is +automatically kept up-to-date. This support is similar to the use_iterator in +the LLVM IR level.</li> + +<li>The MachineInstr, MachineOperand and TargetInstrDesc classes are simpler, +more consistent, and better documented.</li> </ul> </div> <!--=========================================================================--> <div class="doc_subsection"> -<a name="codegen">Code Generator Improvements</a> +<a name="optimizer">Optimizer Improvements</a> </div> <div class="doc_text"> -<p>foci of this release was performance tuning and bug - fixing. In addition to these, several new major changes occurred:</p> +<p>In addition to a huge array of bug fixes and minor performance tweaks, LLVM +2.2 supports a few major enhancements:</p> <ul> -<li>Owen contributed Machine Loop info, domintors, etc. Merged dom and - postdom.</li> +<li>Daniel Berlin and Curtis Dunham rewrote Andersen's alias analysis to be +several orders of magnitude faster, implemented Offline Variable Substitution +and Lazy Cycle Detection. Note that Andersen's is not enabled in llvm-gcc by +default.</li> -<li>Dan added support for emitting debug information with .file and .loc on -targets that support it</li> +<li>Dan Gohman contributed several enhancements to Loop Strength Reduction (LSR) +to make it more aggressive with SSE intrinsics.</li> -<li>Evan physical register dependencies in the BURR scheduler</li> +<li>Evan added support for simple exit value substitution to LSR.</li> -<li>Evan EXTRACT_SUBREG coalescing support</li> <li>Evan enhanced LSR to support induction variable reuse when the induction - variables have different widths.</li> - -<li>Rafael Espindola byval support, finished by Evan for x86 (32/64).</li> +variables have different widths.</li> -<li>mod/ref info for machineinstrs, inferred from .td files</li> - -<li>Evan: Live interval splitting on MBB boundaries.</li> - -Efficient iteration over use/defs of registers like llvm ir. - -The MachineInstr, MachineOperand and TargetInstrDesc classes are simpler, more -consistent, and better documented. </ul> </div> @@ -217,28 +249,11 @@ consistent, and better documented. <li>Bruno: MIPS PIC support.</li> <li>Arnold Schwaighofer: X86 tail call support.</li> <li>Dale darwin/x86-64 and darwin/ppc eh</li> -<li>Evan: darwin/x86 debug info</li> -</ul> - -</div> - - -<!--=========================================================================--> -<div class="doc_subsection"> -<a name="llvmgccimprovements">llvm-gcc Improvements</a> -</div> - -<div class="doc_text"> -<p>New features include: -</p> - -<ul> -<li>.</li> +<li>Evan: darwin/x86 debug info, improvements at -O0?</li> </ul> </div> - <!--=========================================================================--> <div class="doc_subsection"> <a name="coreimprovements">LLVM Core Improvements</a> @@ -260,6 +275,13 @@ consistent, and better documented. <li>Duncan moved parameter attributes off of function type and onto functions and calls, which makes it much easier to add attributes to a function in a transformation.</li> +<li>Christopher Lamb: Multiple address spaces.</li> +<li>Gordon: C and Ocaml Bindings</li> + +<li>Anton added readnone/readonly attributes for modeling function side effects +and Duncan hooked up GCC's pure/const attributes to use them and enhanced alias +analysis to use them.</li> + </ul> </div> |