[MPlayer-dev-eng] [RFC] disable fastmemcpy on x86-64 by default

[Reimar D?ffinger]

Hello,
since SSE is part of the x86-64 architecture, at least glibc makes use
of it for its memcpy and some quick (and imprecise) tests indicate that
it's at least not slower.
So what do you think about attached patch? Can someone do more concise
benchmarks?

Greetings,
Reimar Döffinger
-------------- next part --------------
Index: configure
===================================================================
--- configure	(revision 23390)
+++ configure	(working copy)
@@ -575,7 +576,7 @@
 _sgiaudio=auto
 _sunaudio=auto
 _alsa=auto
-_fastmemcpy=yes
+_fastmemcpy=auto
 _unrarlib=yes
 _win32dll=auto
 _select=yes
@@ -6521,6 +6542,12 @@
 echores "$_mencoder"
 
 echocheck "fastmemcpy"
+if test "$_fastmemcpy" = auto ; then
+  _fastmemcpy=yes
+  # x86-64 default memcpy should be just as good and
+  # can be replaced by a compiler-builtin
+  x86_64 && _fastmemcpy=no
+fi
 # fastmemcpy check is done earlier with tests of CPU & binutils features
 if test "$_fastmemcpy" = yes ; then
   _def_fastmemcpy='#define USE_FASTMEMCPY 1'