Hi,

I compiled POV-Ray with gcc 2.96 including the following flags for Pentium
optimisation as I found in the standard Makefile for POV-Ray:

CFLAGS





= -O6 -finline-functions -ffast-math -c -ansi -m386 -DCPU=686 -DCOMPILER_VER
=\".`uname`.$(CC)\" -DPOV_LIB_DIR=\"$(POVLIBDIR)\" $(SRCINC) $(LIBPNGINC)
$(ZLIBINC)

mixed with optimisations suggested at the athlonlinux.org page:

CFLAGS
= -s -O3 -fomit-frame-pointer -Wall -mpentiumpro -march=pentiumpro -malign-f
unctions=4 -funroll-loops -fexpensive-optimizations -malign-double -fschedul
e-insns2

the result was:

CFLAGS
= -s -O3 -fomit-frame-pointer -Wall -mpentiumpro -march=pentiumpro -malign-f
unctions=4 -funroll-loops -fexpensive-optimizations -malign-double -fschedul
e-insns2 -finline-functions -ffast-math -c -ansi -DCPU=686 -DCOMPILER_VER=\"
.`uname`.$(CC)\" -DPOV_LIB_DIR=\"$(POVLIBDIR)\" $(SRCINC) $(LIBPNGINC)
$(ZLIBINC)

After the compiling the render time for a scene decreased from 7 hours with
the standard POV-Ray binary to 6 hours with my own compiled version. I think




and optimisation things. Maybe there are further optimisation flags for the
Athlon, maybe not.

Is anyone out there who has done an optimised compile for the Athlon too?
Could you share your experience?

Thanks,

Bonsai

Post a reply to this message