On Sunday, June 11, 2017 at 10:13:52 AM UTC-7, David Joyner wrote: > > On Sun, Jun 11, 2017 at 11:30 AM, John H Palmieri > <jhpalm...@gmail.com <javascript:>> wrote: > > > > > > On Saturday, June 10, 2017 at 7:05:49 PM UTC-7, David Joyner wrote: > >> > >> On Thu, Jun 8, 2017 at 10:49 PM, John H Palmieri <jhpalm...@gmail.com> > >> wrote: > >> > > >> > > >> > On Thursday, June 8, 2017 at 5:16:15 PM UTC-7, David Joyner wrote: > >> >> > >> >> Hi all: > >> >> > >> >> This is kind of a newbie question, as I'm not an expert on > simplicial > >> >> complexes. It's also a minor technical issue on the documentation, > not > >> >> the code. > >> >> > >> >> I'm trying to debug some code of mine and, on reading the > >> >> documentation for chain_complex on > >> >> > >> >> > >> >> > http://doc.sagemath.org/html/en/reference/homology/sage/homology/simplicial_complex.html, > > > >> >> found this: "The rows and columns of the boundary matrices are > indexed > >> >> by the lists given by the n_cells() method, which by default are > >> >> sorted." On the other hand, the documentation for n_cells has this > >> >> statement: "Sort the list if the argument sort is True. If sort is > >> >> None (the default), then sort depending on the value of the > >> >> sort_facets parameter (from the initialization of the simplicial > >> >> complex)." However, the sort_facets parameter (unless I'm mistaken) > >> >> does not sort the individual faces but rather sorts the vertices in > >> >> each facet. > >> > > >> > > >> > The documentation means that if "sort_facets" was True originally, > then > >> > n_cells will also sort its result before returning it. ("then sort > >> > depending > >> > on the value of the sort_facets parameter": so if sort_facets was > True, > >> > then > >> > n_cells should also sort. It doesn't mean "sort using the effect of > the > >> > sort_facets parameter", but sort depending on its value, which is > True > >> > or > >> > False. This could be made clearer, I guess.) > >> > > >> > >> > >> To be clear, I am getting different computations using the same sage > >> computations on different machines. One (recent, but used, linux) > >> machine consistently gives ones result, the other gives (randomly?) > >> several others. The "other" (a 8 yr old, a mac with possibly memory > >> issues) machine gives different results. I'm not saying your methods > >> are machine dependent, just that *for me* they seem to be. > >> > >> To be clear, these "inconsistent" computations take at least one day. > >> The "toy" (consistent) examples are okay. > > > > > > That's strange. Are you just getting different orders when you sort, or > > different answers for homology? Are the answers consistent on each > machine? > > What type of object are the vertices? > > > > The example which gives strange output is the bipyramid complex: > > S = SimplicialComplex(maximal_faces=[(1,2,3), (1,2,4),\ > (1,2,5), (1,3,4),(1,3,5),(2,3,4),(2,3,5)]) > > which has 9 1-faces. I'm computing the list of all stable configurations, > namely vectors c in ZZ^9 having non-negative components which cannot > fire (in the sense of chip-firing - see Duval, Klivans, Martin > https://arxiv.org/pdf/1101.3981.pdf) with respect to the 1-dimls > simplicial > spanning tree given by the 1-faces (1, 4), (1,5), (2,5), (3,4). On an old > mac running OS 10.11.6 and sage 7.2.b0, the output is usually 268125, > after about a day, but once I also got 137500. On a linux machine running > ubuntu and sage 7.3, I always get 395625. These computations use > the S.chain_complex.differential method to compute the 1-diml > combinatorial up Laplacian, Q, then does a time-consuming search. > (Basically, I check that no column vector q of Q, except possibly one > corresponding to a face in the spanning tree, has the property that > all components of c-q remain non-negative, except possibly those > corresponding to a face in the spanning tree.) > > A similar computation using the 3-simplex is relatively quick and > always produces > 2500 stable configurations for both machines. >
The sorting was introduced after 7.2, so it's not surprising if that version gives different results. The sorting certainly affects the chain complex, and it is possible in older versions that the ordering used for the simplices to define the matrices for the chain complex is somewhat arbitrary, and not necessarily consistent with other methods for simplicial complexes. I would trust Sage 7.3 over 7.2 because of this sorting. I don't think any relevant changes have been made since 7.3, but it would be interesting to see if you get different results with 7.6 or 8.0. John > > > > > >> > >> > >> > I don't know if there is ever a good reason to specify > >> > sort_facets=False. > >> > Maybe we should deprecate it? > >> > > >> >> > >> >> In my code, I expected the faces of a fixed dimension to > >> >> be returned as a deterministic sorted list not a set whose ordering > >> >> could be random. Now, it *does* seem as though the faces are sorted. > >> >> It's just that, to me, the documentation of n_cells doesn't make > that > >> >> clear. > >> >> > >> >> I'm asking if I'm mis-understanding something here or not. > >> >> > >> >> - David Joyner > >> > > >> > > >> > -- > >> > John > >> > > >> > > >> > -- > >> > You received this message because you are subscribed to the Google > >> > Groups > >> > "sage-support" group. > >> > To unsubscribe from this group and stop receiving emails from it, > send > >> > an > >> > email to sage-support...@googlegroups.com. > >> > To post to this group, send email to sage-s...@googlegroups.com. > >> > Visit this group at https://groups.google.com/group/sage-support. > >> > For more options, visit https://groups.google.com/d/optout. > > > > -- > > You received this message because you are subscribed to the Google > Groups > > "sage-support" group. > > To unsubscribe from this group and stop receiving emails from it, send > an > > email to sage-support...@googlegroups.com <javascript:>. > > To post to this group, send email to sage-s...@googlegroups.com > <javascript:>. > > Visit this group at https://groups.google.com/group/sage-support. > > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "sage-support" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-support+unsubscr...@googlegroups.com. To post to this group, send email to sage-support@googlegroups.com. Visit this group at https://groups.google.com/group/sage-support. For more options, visit https://groups.google.com/d/optout.
