On Wed, Oct 22, 2008 at 8:48 AM, Simon King <[EMAIL PROTECTED]> wrote:
>
> Dear Team,
>
> at SD 10, Martin Albrecht and I implemented the F5 algorithm according
> to John Perry's pseudocode. The two implementations are at
> http://wiki.sagemath.org/days10/CodingSprint
> attachment f5.py (Martin's
On Wednesday 22 October 2008, Simon King wrote:
> Dear Team,
>
> at SD 10, Martin Albrecht and I implemented the F5 algorithm according
> to John Perry's pseudocode. The two implementations are at
> http://wiki.sagemath.org/days10/CodingSprint
> attachment f5.py (Martin's pure python implementatio
On Wed, Oct 22, 2008 at 9:53 AM, Martin Albrecht
<[EMAIL PROTECTED]> wrote:
> The question is: What purpose would such an implementation have:
> (a) educational (i.e. quite read-able/hack-able code)
> (b) coverage (i.e. provide GB calculations for fields Singular doesn't
> support)
Nicolas Thiery
On Wednesday 22 October 2008, Mike Hansen wrote:
> Nicolas Thiery mentioned that F5 works for a class non-commutative
> rings so that might be a reason for including it.
Hi there,
I don't see why F5 would be better suited for non-commutative rings than the
Buchberger (except for speed of course
On Wed, Oct 22, 2008 at 12:53 PM, Martin Albrecht
<[EMAIL PROTECTED]> wrote:
>
> On Wednesday 22 October 2008, Simon King wrote:
>> Dear Team,
>>
>> at SD 10, Martin Albrecht and I implemented the F5 algorithm according
>> to John Perry's pseudocode. The two implementations are at
>> http://wiki.s
I think there is value for development and education in having them
both in. Is it still true that Perry is working on putting a version
in Singular? Even so, if someone improves the cython version it seems
possible that it could become very competitive.
-M. Hampton
On Oct 22, 11:26 am, "David
> Is it still true that Perry is working on putting a version in Singular?
I personally am not writing the code. I did offer, but Christian Eder,
a student at the University of Kaiserslautern, has the primary
responsibility. (I had worked with him on the original toy
implementation as an interpre
On Wed, Oct 22, 2008 at 3:47 PM, john_perry_usm <[EMAIL PROTECTED]> wrote:
>
>> Is it still true that Perry is working on putting a version in Singular?
>
> I personally am not writing the code. I did offer, but Christian Eder,
> a student at the University of Kaiserslautern, has the primary
> res
> That's what I was about to ask. Interesting! How much faster?
>
> William
I think the very first reason is, that the Singular scripting language
is strictly inferior to Python
(this is why I support Sage).
Nevertheless. I just uploaded some nice example to the wiki, for
testing your F5 impl
By the way, I would be interested to know, I they get a result at all
with F5 in this example (this would already be great, but no
pizza ;-) ).
Michael
--~--~-~--~~~---~--~~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this gr
Dear Michael,
On Oct 23, 7:47 am, Michael Brickenstein <[EMAIL PROTECTED]> wrote:
> Nevertheless. I just uploaded some nice example to the wiki, for
> testing your F5 implementation.
> I was originally provided by Gema Diaz.
> It is inhomogeneous and the GB is [1].
>
> http://wiki.sagemath.org/da
Hi Simon!
Am 23.10.2008 um 10:53 schrieb Simon King:
>
> Dear Michael,
>
> On Oct 23, 7:47 am, Michael Brickenstein <[EMAIL PROTECTED]> wrote:
>> Nevertheless. I just uploaded some nice example to the wiki, for
>> testing your F5 implementation.
>> I was originally provided by Gema Diaz.
>> It i
Hi all,
I *really* like the ideas behind slimgb, and I agree with you that
there are more efficiency considerations than reduction to zero. One
of the optimizations I want to try on F5 is a hybrid slimgb/F5
algorithm; I have ideas on how it might work.
I believe that F5 will terminate on the sys
I wrote:
> For example, my toy Singular implementation
> of the basic F5 took nearly a month to compute the GB of Cyclic-8; an
> optimized version took about a week.
That wasn't a very smart thing to say; let me clarify. By "optimized"
version, I still mean a toy (=interpreted, not optimized) ve
Hi there,
to me it seems this thread took a wrong turn (from "toy" to "performance").
Neither my nor Simon's implementation will be competitive by any standards
without massive effort resulting in a complete rewrite. This kind of rewrite
seems futile to me because Christian is working on a S
Martin,
> The right thing to do from my perspective is to ask Christian if he wants
> help. John, can you make Christian aware of this thread?
I sent him a copy of the email I posted. Now that I think of it, it
would be nice to send him a link to this conversation, too.
> Another option is to t
Michael,
I can answer one of your challenges: The non-homogenized system
terminated correctly in about 139 minutes on my home computer, a
1.7GHz P4 w/512MB. I used a sugar variant, having modified Martin's
basic F5 sometime last week. No other changes were made to F5.
I realize this isn't worth
Hi John!
Thanks.
This was one of the systems behaving ugly with incremental strategies.
Considering this, your result is excellent.
Did you check for the result being [1]?
I was referring to "termination in reasonable time".
This makes it a little bit easier to "believe in F5".
Would it be possibl
Michael,
> Did you check for the result being [1]?
Sort of. I did this:
sage: len(G)
<<< 1
Since the output G has length one, it's reasonable to conclude that it
ended with {1}; otherwise there should have been at least 6
polynomials. I didn't think to look at G until afterwards, partly
becaus
On Thu, Oct 23, 2008 at 10:13 AM, john_perry_usm <[EMAIL PROTECTED]> wrote:
>
> Michael,
>
>> Did you check for the result being [1]?
>
> Sort of. I did this:
>
> sage: len(G)
> <<< 1
>
> Since the output G has length one, it's reasonable to conclude that it
> ended with {1}; otherwise there shoul
> Do you consider 139 minutes to be reasonable time for a toy
> implementation?
>
Absolutely, we are in the double exponential case, where differences
are quite big and
benchmarking is doomed.
> Similarly, do you have a "toy" implementation of slimgb we could run
> for comparison? By this I mea
> On Wednesday 22 October 2008, Mike Hansen wrote:
> > Nicolas Thiery mentioned that F5 works for a class non-commutative
> > rings so that might be a reason for including it.
>
> I don't see why F5 would be better suited for non-commutative rings than the
> Buchberger (except for speed of course
Hi everybody,
(sorry for my late replies the university where the conference is at which I'm
attending blocks e-mail)
F4 does not require homogenization and only (IIRC) suggests to use the normal
selection strategy. Matrix-F5 (as described in literature) is quite different
from F4 because it
23 matches
Mail list logo
