#14214: Cythoned homsets
----------------------------------------------+-----------------------------
Reporter: SimonKing | Owner: tbd
Type: enhancement | Status: needs_work
Priority: major | Milestone: sage-5.10
Component: performance | Resolution:
Keywords: Hom, cython, cached method | Work issues: How to store
domain and codomain?
Report Upstream: N/A | Reviewers:
Authors: Simon King | Merged in:
Dependencies: #14159, #12951, #13184 | Stopgaps:
----------------------------------------------+-----------------------------
Comment (by SimonKing):
Replying to [comment:50 nbruin]:
> A very rough implementation:
> {{{
> from sage.misc.lazy_attribute import lazy_attribute
>
> cdef class LazyAttribute:
> cdef public f
> cdef public str __name__
> def __init__(self,f):
> self.f=f
> self.__name__=f.__name__
> def __get__(self,instance,cls):
> if instance is None:
> return lazy_attribute(self.f)
> v=self.f(instance)
> setattr(instance,self.__name__,v)
> return v
> }}}
If I understand correctly, you optimize the case that the instance allows
attribute assignment. But I think that won't be relevant in our use case:
If there is attribute assignment (let's call it "case 1"), then we will
use it during `__init__`, and hence an exiting non-data descriptor will be
overridden.
So, the choice of an implementation of the non-data descriptor will
''only'' matter if there is no attribute assignment available (let's call
it "case 2"). And if I am not mistaken, your code will be very slow in
case 2, and certainly not faster than a @lazy_attribute.
I think the questions to be addressed here are these:
- Do we care about the case that a Homset in case 2? Currently, all
Homsets are case 1, so, we could decide that Homset is a base class, and
only sub-classes allow instantiation.
- If we care about case 2: How can we be as fast as possible?
If I understand correctly, using Cython's "property" would be faster than
a @lazy_attribute in case 2, but it would make it impossible to use
regular attributes in case 1, whence a massive slow-down in case 1 (which
is not acceptable, as case 1 is the normal case).
But can we benefit from a Cython implementation of lazy_attribute so that
it becomes faster in case 2 but still results in a non-data descriptor
that can be overridden by an attribute in case 1?
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/14214#comment:51>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-trac?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
