Patchwork [wwwdocs] libquadmath: Mention in the news, add to changes.html

login
register
mail settings
Submitter Tobias Burnus
Date Nov. 16, 2010, 8:34 p.m.
Message ID <4CE2EAE9.40900@net-b.de>
Download mbox | patch
Permalink /patch/71450/
State New
Headers show

Comments

Tobias Burnus - Nov. 16, 2010, 8:34 p.m.
Hi all, hello Gerald,

the patch adds the libquadmath support to the news section on the main 
GCC page and to the GCC 4.6 release notes.

Do you have comments or suggestions before I install it?

Regarding the library itself: I am currently doing a final bootstrapping 
and will then commit it.

Tobias

Patch

Index: index.html
===================================================================
RCS file: /cvs/gcc/wwwdocs/htdocs/index.html,v
retrieving revision 1.765
diff -u -r1.765 index.html
--- index.html	8 Nov 2010 00:44:28 -0000	1.765
+++ index.html	16 Nov 2010 20:30:46 -0000
@@ -50,6 +50,13 @@ 
 <h2 style="margin-top:0pt;" id="news">News</h2>
 
 <dl class="news">
+<dt>November 16, 2010</dt>
+<dd><a href="gcc-4.6/changes.html">GCC 4.6.0</a> will include the
+<tt>libquadmath</tt> library, which provides quad-precision mathematical
+functions on targets supporting the <tt>__float128</tt> datatype. The
+library is used to provide on such targets the <tt>REAL(16)</tt> type in GNU
+Fortran compiler. The library has been contributed by Fran&ccedil;ois-Xavier
+Coudert.</dd>
 
 <dt>October 1, 2010</dt>
 <dd><a href="gcc-4.4/">GCC 4.4.5</a> has been released.</dd>
Index: gcc-4.6/changes.html
===================================================================
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-4.6/changes.html,v
retrieving revision 1.58
diff -u -r1.58 changes.html
--- gcc-4.6/changes.html	7 Nov 2010 13:48:21 -0000	1.58
+++ gcc-4.6/changes.html	16 Nov 2010 20:30:46 -0000
@@ -38,6 +38,13 @@ 
     generic link-time optimization framework (LTO) introduced
     in <a href="../gcc-4.5/changes.html">GCC 4.5.0</a>.</li>
 
+    <li>GCC now ships with the LGPL-licensed <tt>libquadmath</tt> library,
+    which provides for targets with a <tt>__float128</tt> datatype
+    quad-precision mathematical functions. <tt>__float128</tt> is available
+    for targets on 32-bit x86, x86-64 and Itanium architectures. The
+    <tt>libquadmath</tt> library is automatically build on such targets
+    when building the Fortran compiler.</tt>
+
 </ul>
 
 <h2>General Optimizer Improvements</h2>
@@ -181,6 +188,13 @@ 
 
 <h3 id="fortran">Fortran</h3>
   <ul>
+    <li>On systems supporting the <tt>libquadmath</tt> library, GNU Fortran
+    now also supports a quad-precision, <tt>kind=16</tt> floating-point 
+    data type (<tt>REAL(16)</tt>, <tt>COMPLEX(16)</tt>). As the data type is not
+    fully supported in hardware, calculations might be one to two orders
+    of magnitute slower than with the 4, 8 or 10 bytes floating-point data types.
+    This change does not affect systems which support <tt>REAL(16)</tt> in
+    hardware nor those which do not support  <tt>libquadmath</tt>.</li>
     <li>Much improved compile time for large array constructors.</li>
     <li>Improved diagnostics, especially with
       <code>-fwhole-file</code>.</li>