# This is how I got Radau to work on 64bit Linux (Ubuntu 9.04) and numpy 1.2 and python2.6
# This worked on 3 different machines (all ubuntu 9.04 64 bit), with different architectures AMD and 2 Intel's (L7500, Q9450)
# AMD required changes to 2 codes in numpy (...distutils/cpuinfo.py  and ...distutils/fcompiler/gnu.py)
# the Intel systems I merely had to run codes (cpuinfo.py first) then RADAU worked for me
# the original compile of the 2 scripts must have had a problem when I first installed on intels so merely rerunning scripts worked
# I am using  the AMD Phenom as the example as I had to make changes only to AMD Phenom version
# change to editor of choice (nano,kate,..) I used gedit as I figured novices like me could follow better

# method  changes for AMD Phenom tweak as neccessary
#first run  
	cat /proc/cpuinfo
# example output
	name@comp:~$ cat /proc/cpuinfo
	processor	: 0
	vendor_id	: AuthenticAMD
	cpu family	: 16
	model		: 2
	model name	: AMD Phenom(tm) 9850 Quad-Core Processor
	stepping	: 3
	cpu MHz		: 1250.000   # I used cpu scaling I know there isn't 1.25Ghz Phenom
	cache size	: 512 KB
	physical id	: 0
	siblings	: 4
	core id		: 0
	cpu cores	: 4
	apicid		: 0
	initial apicid	: 0
	fpu		: yes
	fpu_exception	: yes
	cpuid level	: 5
	wp		: yes
	flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good pni monitor cx16 lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs
	bogomips	: 5022.19
	TLB size	: 1024 4K pages
	clflush size	: 64
	cache_alignment	: 64
	address sizes	: 48 bits physical, 48 bits virtual
	power management: ts ttp tm stc 100mhzsteps hwpstate
# to check 'vender id' 'cpu family' etc need to make these consistent in numpy
# Then run 
	/usr/lib/python2.6/dist-packages/numpy/distutils/cpuinfo.py 
# and check if output is consistent with your specs
# If consistent just run 
	/usr/lib/python2.6/dist-packages/numpy/distutils/fcompiler/gnu.py
#  then try to run radau , make sure to delete any earlier radau files for earlier attempts or else you get segmentation fault

#If the cpuinfo.py output is not correct ( as with my AMD Phenem machine) change as follows
	sudo gedit /usr/lib/python2.6/dist-packages/numpy/distutils/cpuinfo.py

#for amd phenom changes around line 145 from
	def _is_AMD64(self):
        	return  self.is_AMD() and self.info[0]['family'] == '15' # make consistent with 'cpu family' value from above
#to (line 163)
	def _is_AMD64(self):
        	return  self.is_AMD() and self.info[0]['cpu family'] == '16'
#Then add def for phenom
	def _is_Phenom(self):
	 	return re.match(r'.*?Phenom\b',
                        self.info[0]['model name']) is not None 

# Since my model number was 2 I also changed the following to keep from my machine being labeled a k6-2
    def _is_AthlonK6_2(self):
        return self._is_AMD() and self.info[0]['model'] == '1'  
# my model number was 2 make sure different from your model number

# save and run (ipython)
# then change gnu.py if necessary I didn't need to for Intel but I did for AMD
	sudo gedit /usr/lib/python2.6/dist-packages/numpy/distutils/fcompiler/gnu.py

# example for phenom (change as appropriate) add under  "def get_flags_arch(self):" (line 180)
# added to both subsections " if gnu_ver >= '3.1.1':"(line 214) and "if gnu_ver >= '3.4':"(line 241) since I wasn't sure
	    elif cpu.is_Phenom():
		marh_opt = '-march=opteron'
# could try march_opt = '-march=Phenom' but not sure if that works  there doesn't seem to be any differences in these options

#save and run 

#Radau should now work or at least it did for me