playing with scipy gh-2823 and https://github.com/EvgeniBurovski/scipy/tree/freeze_memoize

gistfile1.py

import scipy.stats as stats

g = stats.gamma.pdf(8., 3.)

# no consequences for 'regular' distributions
print "\nregular @df=5:", stats.t.pdf(1., df=5)
print "\nregular @df=25:", stats.t.pdf(1., df=25)
print "\nregular @df=5:", stats.t.pdf(1., df=5)

#freeze
t5 = stats.t(df=5)
print "\nfrozen @df=5: ", t5.pdf(1.)
print "\nregular @df=25: ", stats.t.pdf(1., df=25)

# 2nd frozen instance
t35 = stats.t(df=35)
print "\nfrozen @df=5: ", t5.pdf(1.)
print "\nfrozen @df=35: ", t35.pdf(1.)
print "\nfrozen @df=5: ", t5.pdf(1.)


#other distributions are intact
print "previously, gamma was ", g
print "now it is still", stats.gamma.pdf(8., 3.)