[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #42314] Matrix transpose multiply performance
|
From: |
Rik |
|
Subject: |
[Octave-bug-tracker] [bug #42314] Matrix transpose multiply performance regression |
|
Date: |
Thu, 08 May 2014 03:02:18 +0000 |
|
User-agent: |
Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:29.0) Gecko/20100101 Firefox/29.0 |
Follow-up Comment #1, bug #42314 (project octave):
This is not a trick. Since version 3.4 of Octave the parser is supposed to
detect matrix transpose multiply. What it should then do is simply pass the
matrix, along with the transpose flag, to BLAS to perform the multiplication.
This is much faster than performing the transpose in Octave and then
performing the multiplication with BLAS.
This worked in 3.4.X and 3.6.X, but apparently was broken with the 3.8.0
release. There was work done on the parser during that time which I assume
explains this.
I've marked the bug as confirmed and as a regression.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?42314>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/