LAPACK Archives

[Lapack] Fwd: [LAPACK Forum] LAPACKE_?laswp - correct nancheck is no nan


Any input on this?
Thank you
Julie

Begin forwarded message:

From: "LAPACK/ScaLAPACK Development" <julie@Domain.Removed>
Date: October 27, 2015 at 1:13:50 PM PDT
To: julie@Domain.Removed
Subject: [LAPACK Forum] LAPACKE_?laswp - correct nancheck is no nancheck ? : 
post : Approved


User  : dbakshee (dmitry.g.baksheev@Domain.Removed)
Subject       : LAPACKE_?laswp - correct nancheck is no nancheck ?
Thread        : LAPACKE_?laswp - correct nancheck is no nancheck ?
Forum : LAPACK/ScaLAPACK Development ? User Discussion ? Bug report
Mode  : post
IP/Host       : 192.198.151.44 / (n/a)
Actions       : [reply] [quote] [edit] [delete] [info] [pm] [email]
Approved      : Approved
The row swap functions have an option of checking if the input matrix 
contains NaN entries. However, the arguments of the function do not tell the 
row size of the input matrix.

Passing LDA instead of the unknown row size to the NaN checking function is 
not very good idea. The NaN checking function should not check the memory 
that is not part of the matrix.

As an option, the laswp function could have guessed the lower boundary on the 
row size of the matrix by scanning the pivot array first. This option does 
not look appealing too, because the guess may not give the actual row size of 
the matrix. 

Even if the guess is correct, checking full matrix may be redundant. For 
instance, the pivot could describe just one interchange of the first and the 
last row, and the laswp function would have to check for NaNs the whole 
matrix while swapping just only two rows of it, which does not seem a good 
thing to do.

As yet another option, the function could do the NaN check only on the rows 
it interchanges. This would penetrate the NaN check into the computation and 
break the pattern of ?first check, then compute?.

Perhaps, removing the NaN checking option from the LAPACKE_?laswp functions 
is the best option.

What do you think about this?

Thanks
Dmitry
Actions       
[reply]       : 
http://icl.cs.utk.edu/lapack-forum/posting.php?mode=reply&f=13&t=4827
[quote]       : 
http://icl.cs.utk.edu/lapack-forum/posting.php?mode=quote&f=13&p=11587
[edit]        : 
http://icl.cs.utk.edu/lapack-forum/posting.php?mode=edit&f=13&p=11587
[delete]      : 
http://icl.cs.utk.edu/lapack-forum/posting.php?mode=delete&f=13&p=11587
[info]        : 
http://icl.cs.utk.edu/lapack-forum/mcp.php?i=main&mode=post_details&f=13&p=11587
[pm]  : 
http://icl.cs.utk.edu/lapack-forum/ucp.php?i=pm&mode=compose&action=quotepost&p=11587
[email]       : 
http://icl.cs.utk.edu/lapack-forum/memberlist.php?mode=email&u=23122
IP/Host       : 192.198.151.44 / (n/a)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<http://lists.eecs.utk.edu/mailman/private/lapack/attachments/20151027/6e7a40e9/attachment.html>

<Prev in Thread] Current Thread [Next in Thread>
  • [Lapack] Fwd: [LAPACK Forum] LAPACKE_?laswp - correct nancheck is no nancheck ? : post : Approved, ICL <=


For additional information you may use the LAPACK/ScaLAPACK Forum.
Or one of the mailing lists, or