#13589: Controlling C3 to solve once for all the Method Resolution Order issues 
for
category classes
----------------------------------------------------------+-----------------
       Reporter:  nthiery                                 |         Owner:  
nthiery                   
           Type:  defect                                  |        Status:  
needs_review              
       Priority:  major                                   |     Milestone:  
sage-5.11                 
      Component:  categories                              |    Resolution:      
                      
       Keywords:  method resolution order, C3             |   Work issues:      
                      
Report Upstream:  N/A                                     |     Reviewers:  
Simon King, Florent Hivert
        Authors:  Nicolas M. Thiéry                       |     Merged in:      
                      
   Dependencies:  #12894, #12876, #11935, #12895, #10193  |      Stopgaps:      
                      
----------------------------------------------------------+-----------------

Comment (by SimonKing):

 I have cythoned the _cmp_key attribute of `CategoryWithParameters` as
 well. The raw improvement is:

 Without the first two patches:
 {{{
 sage: C = Algebras(FractionField(QQ['x']))
 sage: timeit("if C._cmp_key: del C._cmp_key", number=100000)
 100000 loops, best of 3: 5.38 µs per loop
 }}}

 With all patches:
 {{{
 sage: C = Algebras(FractionField(QQ['x']))
 sage: timeit("if C._cmp_key: del C._cmp_key", number=100000)
 100000 loops, best of 3: 1.28 µs per loop
 }}}

 But I think the startup time is what we are really interested in. My
 starting point is sage-5.11.b3 with these patches applied:
 {{{
 trac_14471_dynamic_class_hash.patch
 trac_14471-review.patch
 trac_14516-crystals_speedup-ts.2.patch
 trac_14722-lazy_import_at_startup-nt.patch
 }}}

 I am giving the total time provided by sage -startuptime, in 5 consecutive
 runs:
 {{{
 Total time (sum over exclusive time): 1485.052ms
 Total time (sum over exclusive time): 1499.287ms
 Total time (sum over exclusive time): 1499.921ms
 Total time (sum over exclusive time): 1492.086ms
 Total time (sum over exclusive time): 1493.375ms
 }}}

 After applying the first two patches, I get:
 {{{
 Total time (sum over exclusive time): 1495.342ms
 Total time (sum over exclusive time): 1504.880ms
 Total time (sum over exclusive time): 1502.199ms
 Total time (sum over exclusive time): 1497.281ms
 Total time (sum over exclusive time): 1508.277ms
 }}}

 And after applying the other two patches, I get:
 {{{
 Total time (sum over exclusive time): 1499.909ms
 Total time (sum over exclusive time): 1498.731ms
 Total time (sum over exclusive time): 1499.289ms
 Total time (sum over exclusive time): 1502.179ms
 Total time (sum over exclusive time): 1489.887ms
 }}}

 I think one sees a tendency. But we'd really need to see what the startup-
 time plugin has to say!

 Apply trac_13589-categories-c3_under_control-nt.patch
 trac13589_cmp_key_attribute.patch trac13589_improve_startuptime.patch
 trac13589_cython_cmp_key.patch

-- 
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/13589#comment:60>
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.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to