|
|
Ublas : |
From: Karl Meerbergen (Karl.Meerbergen_at_[hidden])
Date: 2008-01-14 03:09:12
As far as I remember, norm_1 does not compute the 1-norm for complex
value_type's, unless this bug is now solved (or the documentation modified).
Best,
Karl
Michele De Stefano wrote:
> I answer points 1 and 2. For what regards point 3, I have to say that
> I've never used the scalar matrix ... may be someone else can answer.
>
> 1) In uBLAS, as in any other matrix computational libray, there is no
> function to perform the inverse explicitly. If you need the inverse to
> solve a system, you must exploit numerical solvers (LU decomposition,
> for example). There are some LAPACK bindings around ... I don't
> remember where to download them ... try the "Effective uBLAS" link.
>
> 2) norm_1 performs the computation of the L1 norm of a vector, not
> simply the sum of its elements. The L1 norm, by definition, is exactly
> the sum of absolute values (see Wikipedia or MathWorld).
>
> Michele
>
> 2008/1/8, nchygy <nchygy_at_[hidden] <mailto:nchygy_at_[hidden]>>:
>
> Hello,
> I have several question on using ublas.
> 1. In ublas is it provide any function or method for matrix to
> calculate the inverse?
> 2. In the overview of ublas "norm_1 (x)" has been described to be
> an function to calculate "sum xi" However when i use it calculate
> the sum of an vector x=(87, -32,-34)
> it gives an result 153, that is an wrong number. It seems the
> norm_1 calculate the sum |xi|
> Is it the truth?
> 3. I found it is difficult to use the scalr matrix ,because it can
> not be used to multiply with other matrix.(it show me error, when
> i tring to do so using the function "prod").
> Best regards
> Yang GAO
>
> --
> ¸ßÑï
>
> _______________________________________________
> ublas mailing list
> ublas_at_[hidden] <mailto:ublas_at_[hidden]>
> http://lists.boost.org/mailman/listinfo.cgi/ublas
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>ublas mailing list
>ublas_at_[hidden]
>http://lists.boost.org/mailman/listinfo.cgi/ublas
>
>