Differences between revisions 1 and 12 (spanning 11 versions)
Revision 1 as of 2009-04-09 02:57:07
Size: 639
Editor: was
Comment:
Revision 12 as of 2009-06-05 01:52:13
Size: 3936
Editor: WilliamHart
Comment:
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
= List of Computations where Sage is Noticeably Faster than Magma = = List of Computations where Sage is Noticeably Faster than Magma.... =
Line 3: Line 3:
* Large degree polynomial multiplication modulo n == Machines used ==

eno: (a binary of Sage 4.0.1-rc1 is available at /home/wbhart/sage-4.0.1.rc1/sage on eno)
(a script to stop background processes for benchmarking purposes is available at /home/wbhart/script - but please stop it when done)
Line 5: Line 9:
4-core: model name : Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz
}}}

== Benchmarks ==

* Large degree polynomial multiplication modulo n (sage is three times as fast):
{{{
[[email protected] sage-4.0.1.rc1]$ ./sage
----------------------------------------------------------------------
| Sage Version 4.0.1.rc1, Release Date: 2009-06-04 |
| Type notebook() for the GUI, and license() for information. |
----------------------------------------------------------------------
sage: magma.version()
((2, 15, 8), 'V2.15-8')
Line 8: Line 26:
CPU times: user 0.19 s, sys: 0.00 s, total: 0.19 s
Wall time: 0.19 s
}}}

In my initial tests this seems to be nearly 10 times faster in Sage than in Magma!
{{{
CPU times: user 0.18 s, sys: 0.00 s, total: 0.18 s
Wall time: 0.18 s
Line 17: Line 31:
'Time: 1.690'
sage: 1.69/0.19
8.89473684210526
'Time: 0.540'
Line 21: Line 33:

* Computing factorials (Sage is more than twice the speed).
{{{
[[email protected] sage-4.0.1.rc1]$ ./sage
----------------------------------------------------------------------
| Sage Version 4.0.1.rc1, Release Date: 2009-06-04 |
| Type notebook() for the GUI, and license() for information. |
----------------------------------------------------------------------
sage: magma.version()
((2, 15, 8), 'V2.15-8')
sage: time n = factorial(10^6)
CPU times: user 0.57 s, sys: 0.01 s, total: 0.58 s
Wall time: 0.59 s
sage: time magma.eval('time n := Factorial(10^6);')
CPU times: user 0.00 s, sys: 0.00 s, total: 0.00 s
Wall time: 1.45 s
'Time: 1.440'
sage: time magma.eval('time n := Factorial(10^7);')
CPU times: user 0.00 s, sys: 0.00 s, total: 0.00 s
Wall time: 27.33 s
'Time: 27.300'
sage: time n = factorial(10^7)
CPU times: user 11.50 s, sys: 0.25 s, total: 11.75 s
Wall time: 11.75 s
sage: 27.30/11.75
2.32340425531915
}}}

* Rank of random dense matrices over GF(2)

{{{
----------------------------------------------------------------------
| Sage Version 4.0.alpha0, Release Date: 2009-05-15 |
| Type notebook() for the GUI, and license() for information. |
----------------------------------------------------------------------
sage: A = random_matrix(GF(2),10^4,10^4)
sage: %time A.rank()
CPU times: user 1.23 s, sys: 0.01 s, total: 1.23 s
Wall time: 1.25 s
9998

sage: A = random_matrix(GF(2),2*10^4,2*10^4)
sage: %time A.rank()
CPU times: user 9.65 s, sys: 0.02 s, total: 9.67 s
Wall time: 9.85 s
19937
sage: A = random_matrix(GF(2),2*10^4,2*10^4)
sage: %time A.echelonize(algorithm='pluq')
CPU times: user 7.26 s, sys: 0.02 s, total: 7.27 s
Wall time: 7.40 s

sage: A = random_matrix(GF(2),3.2*10^4,3.2*10^4)
sage: %time A.rank()
CPU times: user 34.60 s, sys: 0.05 s, total: 34.65 s
Wall time: 35.21 s
19937
sage: %time A.echelonize(algorithm='pluq')
CPU times: user 28.35 s, sys: 0.04 s, total: 28.39 s
Wall time: 28.86 s
}}}

{{{
Magma V2.15-8 Sun May 17 2009 13:16:26 on eno [Seed = 595144467]
Type ? for help. Type <Ctrl>-D to quit.
> A:=RandomMatrix(GF(2),10^4,10^4);
> time Rank(A);
10000
Time: 2.790

> A:=RandomMatrix(GF(2),2*10^4,2*10^4);
> time Rank(A);
20000
Time: 19.500

> A:=RandomMatrix(GF(2),32*10^3,32*10^3);
> time Rank(A);
31999
Time: 63.480
}}}


= ....But Magma has the following features which Sage doesn't have (yet) =

* fast and correct multivariate polynomial factorisation algorithm

* fast Gröbner basis computations mod p (p > 2, p prime) and QQ

* fast GCD of multivariate polynomials

List of Computations where Sage is Noticeably Faster than Magma....

Machines used

eno: (a binary of Sage 4.0.1-rc1 is available at /home/wbhart/sage-4.0.1.rc1/sage on eno) (a script to stop background processes for benchmarking purposes is available at /home/wbhart/script - but please stop it when done)

4-core: model name      : Intel(R) Core(TM)2 Quad CPU    Q6600  @ 2.40GHz

Benchmarks

* Large degree polynomial multiplication modulo n (sage is three times as fast):

[[email protected] sage-4.0.1.rc1]$ ./sage
----------------------------------------------------------------------
| Sage Version 4.0.1.rc1, Release Date: 2009-06-04                   |
| Type notebook() for the GUI, and license() for information.        |
----------------------------------------------------------------------
sage: magma.version()
((2, 15, 8), 'V2.15-8')
sage: R.<t> = Zmod(next_prime(8000^3))[]
sage: ff = R.random_element(degree=3200)
sage: time v = [ff*ff for i in [1..100]]
CPU times: user 0.18 s, sys: 0.00 s, total: 0.18 s
Wall time: 0.18 s
sage: S = magma(R)
sage: f = magma('%s![Random(0,10000000) : i in [1..3200]]'%S.name())
sage: magma.eval('time z:=[%s*%s : i in [1..100]]'%(f.name(), f.name()))
'Time: 0.540'

* Computing factorials (Sage is more than twice the speed).

[[email protected] sage-4.0.1.rc1]$ ./sage
----------------------------------------------------------------------
| Sage Version 4.0.1.rc1, Release Date: 2009-06-04                   |
| Type notebook() for the GUI, and license() for information.        |
----------------------------------------------------------------------
sage: magma.version()
((2, 15, 8), 'V2.15-8')
sage: time n = factorial(10^6)
CPU times: user 0.57 s, sys: 0.01 s, total: 0.58 s
Wall time: 0.59 s
sage: time magma.eval('time n := Factorial(10^6);')
CPU times: user 0.00 s, sys: 0.00 s, total: 0.00 s
Wall time: 1.45 s
'Time: 1.440'
sage: time magma.eval('time n := Factorial(10^7);')
CPU times: user 0.00 s, sys: 0.00 s, total: 0.00 s
Wall time: 27.33 s
'Time: 27.300'
sage: time n = factorial(10^7)
CPU times: user 11.50 s, sys: 0.25 s, total: 11.75 s
Wall time: 11.75 s
sage: 27.30/11.75
2.32340425531915

* Rank of random dense matrices over GF(2)

----------------------------------------------------------------------
| Sage Version 4.0.alpha0, Release Date: 2009-05-15                  |
| Type notebook() for the GUI, and license() for information.        |
----------------------------------------------------------------------
sage: A = random_matrix(GF(2),10^4,10^4)
sage: %time A.rank()
CPU times: user 1.23 s, sys: 0.01 s, total: 1.23 s
Wall time: 1.25 s
9998

sage: A = random_matrix(GF(2),2*10^4,2*10^4)
sage: %time A.rank()
CPU times: user 9.65 s, sys: 0.02 s, total: 9.67 s
Wall time: 9.85 s
19937
sage: A = random_matrix(GF(2),2*10^4,2*10^4)
sage: %time A.echelonize(algorithm='pluq')
CPU times: user 7.26 s, sys: 0.02 s, total: 7.27 s
Wall time: 7.40 s

sage: A = random_matrix(GF(2),3.2*10^4,3.2*10^4)
sage: %time A.rank()
CPU times: user 34.60 s, sys: 0.05 s, total: 34.65 s
Wall time: 35.21 s
19937
sage: %time A.echelonize(algorithm='pluq')
CPU times: user 28.35 s, sys: 0.04 s, total: 28.39 s
Wall time: 28.86 s

Magma V2.15-8     Sun May 17 2009 13:16:26 on eno      [Seed = 595144467]
Type ? for help.  Type <Ctrl>-D to quit.
> A:=RandomMatrix(GF(2),10^4,10^4);
> time Rank(A);
10000
Time: 2.790

> A:=RandomMatrix(GF(2),2*10^4,2*10^4);
> time Rank(A);
20000
Time: 19.500

> A:=RandomMatrix(GF(2),32*10^3,32*10^3);
> time Rank(A);
31999
Time: 63.480

....But Magma has the following features which Sage doesn't have (yet)

* fast and correct multivariate polynomial factorisation algorithm

* fast Gröbner basis computations mod p (p > 2, p prime) and QQ

* fast GCD of multivariate polynomials

sagebeatsmagma (last edited 2009-06-12 09:39:03 by was)