#16127: Fix comparison of PARI objects
-------------------------------------+-------------------------------------
Reporter: pbruin | Owner:
Type: defect | Status: new
Priority: critical | Milestone: sage-6.4
Component: interfaces | Resolution:
Keywords: pari comparison | Merged in:
Authors: | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
u/pbruin/16127-gcmp | a119b00099787f5cb8c3cc8477119184e16806ee
Dependencies: | Stopgaps:
-------------------------------------+-------------------------------------
Changes (by pbruin):
* cc: was (added)
Comment:
Here are some timings for this experimental branch:
{{{
a = pari(1)
b = pari(1/2)
c = pari(3.0)
d = pari('x')
e = pari('O(x)')
f = pari('O(2)')
k.<t> = FiniteField(29^10)
u = t^2
Timings with %timeit -c -r 1
cmp cmp == == < <
before after before after before after
(a, a) 189 ns 173 ns 151 ns 144 ns 156 ns 154 ns
(a, b) 1.38 µs 292 ns 568 ns 284 ns 608 ns 208 ns
(a, c) 328 ns 292 ns 171 ns 276 ns 166 ns 204 ns
(a, d) 1.06 µs 22.4 µs 500 ns 312 ns 572 ns 14.4 µs
(a, e) 968 ns 16.6 µs 512 ns 352 ns 456 ns 13.7 µs
(a, f) 1.23 µs 15.6 µs 580 ns 448 ns 532 ns 15.4 µs
(b, b) 175 ns 174 ns 748 ns 153 ns 864 ns 216 ns
(b, c) 3.96 µs 340 ns 1.59 µs 356 ns 1.42 µs 224 ns
(b, d) 1.44 µs 26.4 µs 748 ns 284 ns 696 ns 13.8 µs
(b, e) 1.47 µs 23.2 µs 796 ns 296 ns 624 ns 15.6 µs
(b, f) 1.61 µs 26 µs 824 ns 432 ns 696 ns 21.2 µs
(c, c) 176 ns 178 ns 153 ns 171 ns 140 ns 161 ns
(c, d) 3.56 µs 26 µs 1.19 µs 280 ns 1.33 µs 21.2 µs
(c, e) 3.04 µs 27.6 µs 1.24 µs 248 ns 1.18 µs 19.6 µs
(c, f) 3.12 µs 22.4 µs 1.83 µs 360 ns 1.5 µs 13.7 µs
(d, d) 182 ns 180 ns 556 ns 158 ns 492 ns 13.4 µs
(d, e) 1.96 µs 19.2 µs 548 ns 236 ns 540 ns 21.2 µs
(d, f) 2.28 µs 18.8 µs 672 ns 284 ns 640 ns 22 µs
(e, e) 169 ns 174 ns 496 ns 148 ns 532 ns 23.2 µs
(e, f) 2.16 µs 19.6 µs 620 ns 340 ns 548 ns 17.2 µs
(f, f) 179 ns 166 ns 656 ns 160 ns 652 ns 22.8 µs
(t, t) 166 ns 179 ns 568 ns 588 ns 708 ns 632 ns
(t, u) 1.48 µs 1.36 µs 648 ns 632 ns 648 ns 656 ns
}}}
There is a nice speedup in most cases that do not involve a `PariError`,
including practically all equality testings. However, the slowdown in the
cases where it does occur is probably not acceptable. Maybe it is better
to write a variant of `gcmp_sage()` that calls `gcmp()` and catches errors
using PARI's `pari_CATCH` and `pari_TRY` macros so that we don't have to
raise Python exceptions.
--
Ticket URL: <http://trac.sagemath.org/ticket/16127#comment:5>
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/d/optout.
