[xquery-talk] Matrix Multiplication

Adam Retter adam.retter at googlemail.com
Wed Jan 1 07:33:31 PST 2014


I seriously suggest that you email the XQuery Working Group with your use
case. In particular you should contact Ghislain Fourny and Jonathan Robie,
as both have been working on and gathering use cases for arrays in XQuery
3.1. So far efficient matrix operations is not one that I have seen.

Cheers Adam.
On 31 Dec 2013 17:03, "jean-marc Mercier" <jeanmarc.mercier at gmail.com>
wrote:

> @David pairs are also basically needed to write a linear algebra modulus,
> the topic of this thread. And XQUERY don't provide any efficient pair. You
> can't use Marklogic map, or any other vendor map to store vectors for
> performance issues (a map is really slow).
>
> Note that there are a lot of workaround : I am using direct JAVA binding
> or C++ binding from XQUERY for linear algebra not to pay a too heavy
> tribute to these issue performances.
>
> The point is simply to notice that XQUERY could be really good to write an
> efficient linear algebra modulus. But, due to these performance issues,
> nobody can write it. I just hope that the next XQUERY version will give the
> necessary container to write it. Meanwhile, nobody can contribute to XQUERY
> through external modulus using heavy algorithms, and that's just too bad.
>
>
>
>
>
>
> 2013/12/31 David Lee <dlee at calldei.com>
>
>>  Indeed  ... pair( (1,2,3) ) is just a function that returns a function
>> as per my example.
>>
>> But to the point,  you can either use vendor extensions (such as
>> marklogic's  json:array and map:map types) which have good support for
>> efficient operations, or look to another language (<sigh>)  You may find
>> Scala more amenable to this type of programming.
>>
>>
>>
>>
>>
>>
>>
>> *From:* talk-bounces at x-query.com [mailto:talk-bounces at x-query.com] *On
>> Behalf Of *jean-marc Mercier
>> *Sent:* Tuesday, December 31, 2013 11:49 AM
>> *To:* Michael Sokolov
>> *Cc:* xquery-discuss; ihe onwuka; Andrew Welch
>>
>> *Subject:* Re: [xquery-talk] Matrix Multiplication
>>
>>
>>
>> @David, there are tricks to overcome sequence concatenations now. See for
>> instance definition of a pair by Leo, John Snelson, or me : you can write
>> pair( ( 1 , 2 , 3 ) , (4  5 , 6) ) to avoid sequence concatenation. I ve
>> written also constant sized vectors using this trick : for instance
>> NUPLE(1,(),<toto>,5), withh associated getters.
>>
>> The bad new is that these operations takes too much time with all the
>> interpreters I have tried, and can't be used in heavy algorithms.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> 2013/12/31 jean-marc Mercier <jeanmarc.mercier at gmail.com>
>>
>> Hi
>>
>>
>>
>> @Michael Concerning your remark over the discussions I quoted, maps are
>> the basic block for sparse linear algebra. Without performent "maps of
>> nodes" (equivalent to std::map in C++) you will not be able to write any
>> performant linear algebra modulus for sparse matrix.
>>
>>
>>
>> However, before even thinking to sparse matrix, operations on sequences
>> as concatenations are the first "show-stop" to write a linear algebra
>> modulus, since sequences are vectors.
>>
>> Another one is the lack of constant sized vectors (needed for basic dense
>> matrix operations).
>>
>>
>>
>> @Ryan thx for these links, they are very interesting.
>>
>>
>>
>> Well, I am going to party ! Have a happy new year !
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> 2013/12/31 Michael Sokolov <msokolov at safaribooksonline.com>
>>
>> I would love to see some tests of pure XQuery implementations of both
>> sparse and dense operations.  I suspect performance of matrix multiply,
>> inversion, etc, will be poorer than in C++ or Java, but I would expect
>> performance comparable to Perl or Python (w/o its numpy extension) - just a
>> wild guess.  I'd also expect it might be easier to get good sparse
>> performance.  I don't know why, just a hunch.
>>
>> But the more interesting question for me is whether language changes are
>> really needed to support this.  I would have thought that proper
>> optimization of operations on sequences would be enough?  So for example,
>> an extension module using sequences as matrix datatypes could possibly
>> optimize performance using a lower-level implementation.  Does anyone see
>> any reason why that wouldn't be possible?
>>
>> -Mike
>>
>> PS:
>> I reviewed the discussion you referred to, jean-marc, but it seems to
>> have more to do with using functions as map keys, and I don't see any
>> direct connection to linear algebra?
>>
>>
>>
>> On 12/31/2013 9:55 AM, jean-marc Mercier wrote:
>>
>>   It is not due to the spec. It is rather due to the common usage of
>> XQUERY, forcing vendor solutions (as BaseX) to focus primarily on XML Data
>> Base requests more than algorithmic performances.
>>
>>
>>
>> There are actually some threads that are discussing these performance
>> issues in the context of maps (maps are for instance used for sparse matrix
>> representations) : look for instance to ""map module for XQUERY ?" that I
>> initiated or "Higher-order XQuery Modules" by Leo from BaseX, on talk@
>> x-query.com mailing list.
>>
>> Anyhow, to write a serious linear algebra modulus, the basic need is to
>> have a vector containers. Unfortunately, XQUERY does not provide any
>> performant vector containers at present time, and it is not possible to
>> code them in pure XQUERY : I have tried, and more experienced xquery
>> developpers than me have also tried, without success.
>>
>>
>>
>> We have to wait for the XQUERY version that will give us these
>> containers, a decision to be taken by the W3C.
>>
>>
>>
>>
>>
>> 2013/12/31 Andrew Welch <andrew.j.welch at gmail.com>
>>
>>
>> Are you saying the spec as it stands somehow forces all implementations
>> to be 1000x slower, or just what you have observed in some particular
>> implementation?
>>
>>   On 31 Dec 2013 14:27, "jean-marc Mercier" <jeanmarc.mercier at gmail.com>
>> wrote:
>> >
>> > As far as I understand, you want to write a linear algebra module using
>> XQUERY ?
>> > If so, I opened a thread some months ago about this idea. My opinion
>> today is that this is a false good idea at present time.
>> >
>> > 1) XQUERY would be really good for writing concise, efficient linear
>> algebra modulus.
>> > 2) However, I strongly recommend to wait a little bit for starting
>> coding : the current version of XQUERY (3.0) suffers from performance
>> issues. A linear algebra modulus written in XQUERY is expected to have
>> performances performances 1000 X slower than its corresponding C++ or JAVA
>> (you can measure it precisely). Any mathematician linear algebra modulus
>> would probably trashed your modulus after the first test.
>> >
>> > Hope this helps
>> >
>> >
>> >
>> > 2013/12/31 Ihe Onwuka <ihe.onwuka at gmail.com>
>> >>
>> >> Assuming a sparse representation it is about 4 lines of SQL. This is
>> known not least because you can read enough articles and papers that
>> discuss it and it optimises well. The obvious google search does not reveal
>> any corresponding XQuery discussion, neither does it appear to have
>> surfaced on this or the eXist mailing list (allowing for my deficient
>> search skills). For something so "trivial" I thought that was rather
>> strange. Hence I thought it would be prudent to ask  before naively
>> embarking on a  600k X 40k matrix multiplication.
>> >>
>> >>
>> >>
>> >> On Tue, Dec 31, 2013 at 11:31 AM, Andrew Welch <
>> andrew.j.welch at gmail.com> wrote:
>> >>>
>> >>>
>> >>> It should be pretty trivial...
>> >>>
>> >>> On 31 Dec 2013 11:07, "Ihe Onwuka" <ihe.onwuka at gmail.com> wrote:
>> >>> >
>> >>> > Has anybody tried this in XQuery or if I am so foolish (not yet but
>> give me time) would I be the courageous  <culturalReference>
>> http://www.youtube.com/watch?v=ik8JT2S-kBE</culturalReference> early
>> adopter.
>> >>> >
>> >>> >
>> >>> > _______________________________________________
>> >>> > talk at x-query.com
>> >>> > http://x-query.com/mailman/listinfo/talk
>> >>
>> >>
>> >>
>> >> _______________________________________________
>> >> talk at x-query.com
>> >> http://x-query.com/mailman/listinfo/talk
>> >
>> >
>>
>>
>>
>>
>>
>> _______________________________________________
>>
>> talk at x-query.com
>>
>> http://x-query.com/mailman/listinfo/talk
>>
>>
>>
>>
>>
>>
>>
>
>
> _______________________________________________
> talk at x-query.com
> http://x-query.com/mailman/listinfo/talk
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://x-query.com/pipermail/talk/attachments/20140101/1ec83887/attachment.html>


More information about the talk mailing list