On 2015-04-20, Simon King <simon.k...@uni-jena.de> wrote:
> Thank you! Is there a penalty for doing "type(self)" twice? I am
> thinking of PY_NEW(type(self)) versus type(self).__new__(type(self)),
> with typical application in the creation of elements in methods like
> _add_ or _mul_.
There is a penalty, but one can cope with it:
from sage.ext.memory cimport *
def test1(C):
cdef size_t n
for n from 0<=n<1000000:
x = PY_NEW(type(C))
def test2(C):
cdef size_t n
for n from 0<=n<1000000:
x = (type(C)).__new__(type(C))
def test1b(C):
cdef size_t n
cdef type T
for n from 0<=n<1000000:
T = type(C)
x = PY_NEW(T)
def test2b(C):
cdef size_t n
cdef type T
for n from 0<=n<1000000:
T = type(C)
x = T.__new__(T)
and we obtain (with a cythoned version of path algebra elements):
sage: DiGraph({0:{1:['a'], 2:['b']}, 1:{0:['c'], 1:['d']},
2:{0:['e'],2:['f']}}).path_semigroup().algebra(ZZ).inject_variables()
Defining e_0, e_1, e_2, a, b, c, d, e, f
sage: %timeit test1(a)
10 loops, best of 3: 101 ms per loop
sage: %timeit test1b(a)
10 loops, best of 3: 105 ms per loop
sage: %timeit test2(a)
10 loops, best of 3: 196 ms per loop
sage: %timeit test2b(a)
10 loops, best of 3: 102 ms per loop
Hence, doing (type(C)).__new__(type(C)) is clearly slower than
PY_NEW(type(C)), but when one assigns type(C) to a local variable, there
is no difference. The only advantage of PY_NEW is that one has less to
write.
Best regards,
Simon
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
