Add blurb about module flags and reformat a bit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150677 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Bill Wendling 2012-02-16 10:23:43 +00:00
parent 6afa3f744a
commit 190ec9b9c1

View File

@ -286,9 +286,8 @@ Release Notes</a>.</h1>
<ul> <ul>
<li>IR support for half float</li> <li>IR support for half float</li>
<li>IR support for vectors of pointers, including vector GEPs.</li> <li>IR support for vectors of pointers, including vector GEPs.</li>
<li>The <tt>unwind</tt> instruction is now gone. With the introduction of <li>Module flags have been introduced. They convey information about the
the new exception handling system in LLVM 3.0, the <tt>unwind</tt> module as a whole to LLVM subsystems.</li>
instruction became obsolete.</li>
<li>....</li> <li>....</li>
</ul> </ul>
</div> </div>
@ -337,12 +336,12 @@ Release Notes</a>.</h1>
<div> <div>
<p>We have changed the way that the Type Legalizer legalizes vectors. <p>We have changed the way that the Type Legalizer legalizes vectors. The type
The type legalizer now attempts to promote integer elements. legalizer now attempts to promote integer elements. This enabled the
This enabled the implementation of vector-select. implementation of vector-select. Additionally, we see a performance boost on
Additionally, we see a performance boost on workloads which use vectors of chars and shorts, since they are now promoted workloads which use vectors of chars and shorts, since they are now promoted
to 32-bit types, which are better supported by the SIMD instruction set. to 32-bit types, which are better supported by the SIMD instruction set.
Floating point types are still widened as before.</p> Floating point types are still widened as before.</p>
<p>We have put a significant amount of work into the code generator <p>We have put a significant amount of work into the code generator
@ -351,18 +350,19 @@ Release Notes</a>.</h1>
<ul> <ul>
<li>TableGen can now synthesize register classes that are only needed to <li>TableGen can now synthesize register classes that are only needed to
represent combinations of constraints from instructions and sub-registers. represent combinations of constraints from instructions and sub-registers.
The synthetic register classes inherit most of their properties form their The synthetic register classes inherit most of their properties form their
closest user-defined super-class.</li> closest user-defined super-class.</li>
<li><code>MachineRegisterInfo</code> now allows the reserved registers to be <li><code>MachineRegisterInfo</code> now allows the reserved registers to be
frozen when register allocation starts. Target hooks should use the frozen when register allocation starts. Target hooks should use the
<code>MRI->canReserveReg(FramePtr)</code> method to avoid accidentally <code>MRI->canReserveReg(FramePtr)</code> method to avoid accidentally
disabling frame pointer elimination during register allocation.</li> disabling frame pointer elimination during register allocation.</li>
<li>A new kind of <code>MachineOperand</code> provides a compact <li>A new kind of <code>MachineOperand</code> provides a compact
representation of large clobber lists on call instructions. The register representation of large clobber lists on call instructions. The register
mask operand references a bit mask of preserved registers. Everything else is mask operand references a bit mask of preserved registers. Everything else
clobbered.</li> is clobbered.</li>
</ul> </ul>
</div> </div>
<!--=========================================================================--> <!--=========================================================================-->
@ -439,9 +439,12 @@ Release Notes</a>.</h1>
from the previous release.</p> from the previous release.</p>
<ul> <ul>
<li>LLVM 3.1 removes support for reading LLVM 2.9 bitcode files. Going forward, <li>LLVM 3.1 removes support for reading LLVM 2.9 bitcode files. Going
we aim for all future versions of LLVM to read bitcode files and .ll files forward, we aim for all future versions of LLVM to read bitcode files and
produced by LLVM 3.0 and later.</li> <tt>.ll</tt> files produced by LLVM 3.0 and later.</li>
<li>The <tt>unwind</tt> instruction is now gone. With the introduction of the
new exception handling system in LLVM 3.0, the <tt>unwind</tt> instruction
became obsolete.</li>
<li>....</li> <li>....</li>
</ul> </ul>