| Ticket UUID: | 1784637 | |||
| Title: | Changes in matmul between vers. 1.0 and 1.0.1 | |||
| Type: | Bug | Version: | None | |
| Submitter: | bwallner | Created on: | 2007-08-30 07:56:12 | |
| Subsystem: | math :: linearalgebra | Assigned To: | arjenmarkus | |
| Priority: | 5 Medium | Severity: | ||
| Status: | Closed | Last Modified: | 2009-11-16 15:22:07 | |
| Resolution: | Fixed | Closed By: | arjenmarkus | |
| Closed on: | 2009-11-16 08:22:07 | |||
| Description: |
I found a different (wrong) behaviuor between version 1.0 and 1.0.1 in matrix multiplikation. See the attached file. The result for variable pvv should be: 1.6848 (vers. 1.0) but I got: 0.5220 0.1606 0.2770 0.3830 0.4059 ... | |||
| User Comments: |
arjenmarkus added on 2009-11-16 15:22:07:
allow_comments - 1 Closing this issue - solved in version 1.0.2 arjenmarkus added on 2007-09-07 02:19:40: Logged In: YES user_id=400048 Originator: NO Yes, the result of matmul_vm (the underlying routine for matmul for the multiplication of a row vector and a matrix) actually returned a column vector. (The change from 1.0 to 1.0.1 was meant to solve the issue of row and column vectors, there is a natural distintion possible between them. I tried to keep some of the previous, sloppy behaviour, while at the same time being more strict in this distinction. I solved the problem by transposing the result - this gives a correct row vector and therefore a correct answer with the rest of the computations) See the CVS repository (version now: 1.0.2) arjenmarkus added on 2007-08-30 16:00:14: Logged In: YES user_id=400048 Originator: NO Oh, I will have to look into that - I do not remember the details of why I made this change. bwallner added on 2007-08-30 14:56:13: File Added - 243494: linalgTest.tcl | |||
Attachments:
- linalgTest.tcl [download] added by bwallner on 2007-08-30 14:56:12. [details]