On 4/21/06, Brian Dunn <job17and9(a)sbcglobal.net> wrote:
I'm running gentoo, and upon reading the gcc flag
descriptions i've
determined that my cpu can support the following CFLAGS
CFLAGS="-O2 -march=pentium4 -pipe -msse -msse2 -mmmx -mfpmath=sse"
cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 2.66GHz
stepping : 9
cpu MHz : 2658.181
cache size : 512 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr
bogomips : 5323.63
and so that's what i've been compiling with. trouble is, rosegarden
doesn't run at all before segfaulting, MusE does pretty well but sort of
segfaults at random. so does anyone think that this instability could
be related to overly agressive cflags?
Probably not. You're CFLAGS, while optimal, are not all that overly
agressive, and the people who write the ebuild scripts usually have
the scripts filter out CFLAG options that are known to cause problems.
Both these projects are a little unstable. (I personally don't have as
much trouble as you are describing, but I have found Muse a bit more
stable than rosegarden)
BTW: you can get rid of -mmmx -msse and -msse2. These are implied by
-march=pentium4. You might also consider adding -ffast-math. Make your
CXXFLAGS the same as your CFLAGS, but add -fvisibility-inlines-hidden.
i think that the fmpath=sse might be a bad idea, as it
sortof forces the
compiler to use the floating point co-processor (?).
in a more genral sense, will all my bells and whistles be used if i
don't set these compile flags, and will they really make that much of a
difference? how about xrun wise?
Does anyone out there have good performace for audo apps with a simalar
chip and what flags are you using?
Thanks ahead of time, and if all this is already discussed somewhere, a
link would be just dandy.
Brian