Then again, if one works in a number field in Magma instead of a
quotient ring, it takes about a quarter of the time of SAGE to do the
power. That's probably about comparable to what we'd get with a single
scaling before and after doing the powering in SAGE. So it looks like
when all the appropria
Magma also takes 0.12s to do the polynomial expmod a^2 mod f, but
when you raise the power to 20 or 200 Magma is not competitive
any more. This is quite surprising, since as everyone knows, Magma is
usually quite competitive with polynomial arithmetic.
Bill.
On 3 Dec, 18:26, Bill Har
On Dec 3, 2007 3:42 PM, William Stein <[EMAIL PROTECTED]> wrote:
> That won't work because nobody ever implement _maple_init_ as a method
> for Sage matrices.
Hmm, I'd be interested in looking into improving this myself.
> > In any case, there's nothing syntactically tricky about matrix
> > exp
On Dec 3, 2007 12:30 PM, Stephen Forrest <[EMAIL PROTECTED]> wrote:
>
> On Dec 3, 2007 11:16 AM, William Stein <[EMAIL PROTECTED]> wrote:
> [snip]
> > I want to time maple, but I just spent 10 minutes and couldn't
> > even figure out how to raise a matrix to a power!
>
> I couldn't get the Sage in
On Dec 3, 2007 11:16 AM, William Stein <[EMAIL PROTECTED]> wrote:
[snip]
> I want to time maple, but I just spent 10 minutes and couldn't
> even figure out how to raise a matrix to a power!
I couldn't get the Sage interface to Maple to work properly with
matrices. When I tried the following in S
On Dec 3, 2007 10:26 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
>
>
>
> On 3 Dec, 16:16, "William Stein" <[EMAIL PROTECTED]> wrote:
>
> > By the way, PARI is *not* competitive here:
> >
> > sage: nn = gp(n)
> > sage: gp.eval('gettime; a = %s^2; gettime/1000.0'%nn.name())
> > '0.42800
On 3 Dec, 16:16, "William Stein" <[EMAIL PROTECTED]> wrote:
> By the way, PARI is *not* competitive here:
>
> sage: nn = gp(n)
> sage: gp.eval('gettime; a = %s^2; gettime/1000.0'%nn.name())
> '0.428000'
> sage: gp.eval('gettime; a = %s^20; gettime/1000.0'
On 3 Dec, 16:16, "William Stein" <[EMAIL PROTECTED]> wrote:
> Where are you getting the above timings from?
I worked in QQ['x'].quotient(f) which is slower than working in a
number field.
Bill.
--~--~-~--~~~---~--~~
To post to this group, send email to sage-deve
On my system, for this example sage takes .36 seconds, and
mathematica takes 3.86 seconds, so sage is about 10 times faster.
-MH
On Dec 3, 10:42 am, Jason Grout <[EMAIL PROTECTED]> wrote:
> William Stein wrote:
> > On Dec 3, 2007 8:13 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
> >> I did try to c
On Dec 3, 2007 8:42 AM, Jason Grout <[EMAIL PROTECTED]> wrot
>
>
> William Stein wrote:
> > On Dec 3, 2007 8:13 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
> >> I did try to check that Mathematica was getting the right answer, but
> >> I had no luck. I don't know how to convert a mathematica matrix i
William Stein wrote:
> On Dec 3, 2007 8:13 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
>> I did try to check that Mathematica was getting the right answer, but
>> I had no luck. I don't know how to convert a mathematica matrix into
>> ordinary matrix form in SAGE, so when I do the comparison it alway
On Dec 3, 2007, at 11:20 AM, William Stein wrote:
>
> On Dec 3, 2007 8:13 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
>> I did try to check that Mathematica was getting the right answer, but
>> I had no luck. I don't know how to convert a mathematica matrix into
>> ordinary matrix form in SAGE, so
On Dec 3, 2007 8:13 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
> I did try to check that Mathematica was getting the right answer, but
> I had no luck. I don't know how to convert a mathematica matrix into
> ordinary matrix form in SAGE, so when I do the comparison it always
> just says false.
Damn
On Dec 3, 2007 5:40 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
>
> I've just been looking at SAGE ticket number 173:
>
> http://www.sagemath.org:9002/sage_trac/ticket/173
>
> The idea is that Mathematica raises a 3 dimensional matrix M over QQ
> to the power 20,000 much faster than either SAGE or Ma
I did try to check that Mathematica was getting the right answer, but
I had no luck. I don't know how to convert a mathematica matrix into
ordinary matrix form in SAGE, so when I do the comparison it always
just says false.
Bill.
On 3 Dec, 15:21, Clement Pernet <[EMAIL PROTECTED]> wrote:
> Hi th
Hi there,
The method using x^k mod charpoly (or minpoly) is clearly the only
method I know about for that problem.
If n is smallish, this is the good way to do it.
For larger n (say n=O(k)), the computation of the n power of A in step 3
is the bottleneck (n^4 or n^(w+1) ops in Q, so roughly O(
On Dec 3, 2007, at 8:40 AM, Bill Hart wrote:
> I've just been looking at SAGE ticket number 173:
>
> http://www.sagemath.org:9002/sage_trac/ticket/173
>
> The idea is that Mathematica raises a 3 dimensional matrix M over QQ
> to the power 20,000 much faster than either SAGE or Magma.
>
> I don't
17 matches
Mail list logo