Post Reply 
Version 6030 and SVD
06-05-2014, 12:49 AM
Post: #1
Version 6030 and SVD
Is singular value decomposition (SVD) broken in version 6030?

I just tried a 3x4 and a 2x3 matix (and their transposes) and the Prime freezes up - - after a while, it will reboot. Aargh!

Try it in Home or CAS.

I believe this worked just fine in the previous firmware.
Find all posts by this user
Quote this message in a reply
06-05-2014, 04:20 AM
Post: #2
RE: Version 6030 and SVD
(06-05-2014 12:49 AM)Helge Gabert Wrote:  Is singular value decomposition (SVD) broken in version 6030?

I just tried a 3x4 and a 2x3 matix (and their transposes) and the Prime freezes up - - after a while, it will reboot. Aargh!

Try it in Home or CAS.

I believe this worked just fine in the previous firmware.

Yup, just did a 3x4 in the Home screen and had to put the pin in it after a couple of minutes.

HP Prime - sexy hardware ... just don't try and confuse it with complicated stuff like mathematics.
Find all posts by this user
Quote this message in a reply
06-05-2014, 07:41 AM (This post was last modified: 06-05-2014 07:43 AM by parisse.)
Post: #3
RE: Version 6030 and SVD
Ok, I've fixed it. Could not see it on giac because lapack is used there.
With the current firmware, SVD will not work for matrices that have more columns than rows, because it does not call correctly qr for completing the basis of q.
norlesh, I'm not convinced you really understand what happens here.
Find all posts by this user
Quote this message in a reply
06-05-2014, 10:13 AM
Post: #4
RE: Version 6030 and SVD
(06-05-2014 07:41 AM)parisse Wrote:  norlesh, I'm not convinced you really understand what happens here.

Parisse, care to elaborate?
Find all posts by this user
Quote this message in a reply
06-05-2014, 04:33 PM
Post: #5
RE: Version 6030 and SVD
(06-05-2014 07:41 AM)parisse Wrote:  Ok, I've fixed it. Could not see it on giac because lapack is used there.
With the current firmware, SVD will not work for matrices that have more columns than rows, because it does not call correctly qr for completing the basis of q.
norlesh, I'm not convinced you really understand what happens here.

Thanks for fixing! Hopefully, with the next firmware update (should there be one), also nxm matrices with m>n can be handled. But all of this was correct in ver 5447, or not?
Find all posts by this user
Quote this message in a reply
06-05-2014, 06:16 PM
Post: #6
RE: Version 6030 and SVD
Maybe, I don't remember, in that case it was perhaps a side effect of improving some commands to handle complex matrices.
Find all posts by this user
Quote this message in a reply
Post Reply 




User(s) browsing this thread: 1 Guest(s)