Re: some unapplied patches

[David Bateman]

John W. Eaton wrote:
> On 12-Feb-2008, David Bateman wrote:
>
> | Here are two patches that haven't been applied yet.. What was the status
> | on them. I modified the lwork calculation for xGELSD such that the value
> | is only set for overdetermined matrices, limiting the effect of this
> | change..
>
> I decided to use Jason's patch for the xGELSD bug.
>   
I think there is a factor of 1.6 (ie M / MNTHR in xGELSD.f) missing in
Jason's patch. But apart from that I'm fine with his patch.. The result
of the missing factor of 1.6 is that the "m*m" memory allocation can end
up being upto the size of the original matrix rather than 1.6^2 smaller
than the original matrix.. So there is a potential of allocating upto 
2.56 times more memory than needed.

D.




-- 
David Bateman                                address@hidden
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 6 72 01 06 33 (Mob) 
91193 Gif-Sur-Yvette FRANCE                  +33 1 69 35 77 01 (Fax) 

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary