I was recompiling my MMX code on GCC 4 and was surprised to see that none of
the moves from general regs to MMX (and other way around) were happening
directly but through intermediate memory location and I could not find a
combination of optimization options which would disable this behaviour.
I tried to find out what is causing this and this is what I found:
gcc/config/i386/i386.c:
int ix86_secondary_memory_needed (enum reg_class class1, enum reg_class
class2, enum machine_mode mode, int strict)
...
/* ??? This is a lie. We do have moves between mmx/general, and for mmx/sse2.
But by saying we need secondary memory we discourage the register allocator
from using the mmx registers unless needed. */
if (MMX_CLASS_P (class1) != MMX_CLASS_P (class2)) return true;
...
If I comment out that "if" line, mmx register assignments from general
registers will work.
Should the requirement of using secondary memory to move MMX registers between
general regs be relaxed ?
I do not see the point why one should discourage the register allocator from
using mmx registers, move through memory is clearly inefficent and enlarges
resulting code (if the function containing moves is inlined in several places,
even more so).
I'll attach my source and assembly output from GCC 4.1.0 (30.08.2005 CVS HEAD)
vs GCC 3.4.4, these are short.
--
Summary: Moves from general to MMX registers use unnecessary
secondary memory
Product: gcc
Version: 4.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: vahur dot sinijarv at cydonia dot ee
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i686-pc-mingw32
GCC host triplet: i686-pc-mingw32
GCC target triplet: i686-pc-mingw32
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23660
