[with patch, positive review] speed problem when factoring polynoms

[Harald Schilly]

There is a huge speed difference. Any special reasons? A novice user would possibly not understand why!

var('x,y')
time p1=factor(x^99+y^99)

Time: CPU 0.05 s, Wall: 58.43 s

R.<x,y> = QQ[]
time p2=factor(x^99+y^99)

Time: CPU 0.06 s, Wall: 0.06 s

[Michael Abshoff]

This is likely caused by using Maxima's factoring vs. Singular's libfactor. In the first case x and y are symbolic.

Cheers,

Michael

[Martin Albrecht]

I vote for wontfix because I see no way of fixing this, this is a Maxima speed issue. The only fix I could think of is to add something about this in some documentation.

[Michael Abshoff]

At least one comment is wrong in the patch since the "-" no longer shows up:

2288	2288	        Notice that the -1 factor is separated out: 
2289	2289	            sage: f.factor_list() 
2290	 	            [(-1, 1), (y - x, 1), (y^2 + x*y + x^2, 1)] 
 	2290	            [(x - y, 1), (y^2 + x*y + x^2, 1)] 

Cheers,

Michael

[ncalexan]

This patch could interact with #1391 (​http://trac.sagemath.org/sage_trac/ticket/1391). That one should be applied first, I think, and then this looked at again.

[Martin Albrecht]

The patch applies to 2.10.1.alpha1 (hunks, but success). Afterwards, only the toy_buchberger.py tests fail which is unrelated. Thus, I say: apply.

[William Stein]

The mabshoff comment above about "At least one comment is wrong in the patch since the "-" no longer shows up" was caused by ncalexan's patch related to factorization.py, which was after #1617.

[Michael Abshoff]

Merged in Sage 2.10.1.alpha2