incubator-hama-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Edward J. Yoon" <edwardy...@apache.org>
Subject Re: 2D blocking algorithm
Date Tue, 21 Oct 2008 01:38:40 GMT
Oh....Sorry, It's really difficult to explain it in English. : (
I'll add it to wiki.

On Tue, Oct 21, 2008 at 10:36 AM, Edward J. Yoon <edwardyoon@apache.org> wrote:
> For example,
> We have a 4 * 4 Matrix a :
>
> [a, b, c, d]
> [e, f, g, h]
> [i, j, k, l]
> [m, n, o, p]]
>
> Then, we can store an 2 * 2 block matrices to file.
>
> block(0, 0) is a 2 * 2 sub-matrix as below.
>
> [a, b]
> [e, f]
>
> Key : block(0, 0), Value : submatrix [a, b], [e, f]
> Key : block(0, 1), Value : submatrix [c, d], [g, h]
> Key : block(1, 0), Value : submatrix [i, j], [m, n]
> Key : block(1, 1), Value : submatrix [k, l], [o, p]
>
> .....
>
> Also, If we can access sub-matrix from Hbase table, blocking can be
> pre-computed.
>
> On Tue, Oct 21, 2008 at 10:18 AM, Samuel Guo <guosijie@gmail.com> wrote:
>> should the matrices that do the multiplication have the same blockID layout?
>>
>> And How to divide a matrix into BlockIDs to make the parallel computation
>> most efficiently?
>>
>> 2008/10/21 Edward J. Yoon <edwardyoon@apache.org>
>>
>>> Can anyone review this ??
>>>
>>> http://blog.udanax.org/2008/10/parallel-matrix-multiply-on-hadoop.html
>>>
>>> --
>>> Best regards, Edward J. Yoon
>>> edwardyoon@apache.org
>>> http://blog.udanax.org
>>>
>>
>
>
>
> --
> Best regards, Edward J. Yoon
> edwardyoon@apache.org
> http://blog.udanax.org
>



-- 
Best regards, Edward J. Yoon
edwardyoon@apache.org
http://blog.udanax.org

Mime
View raw message