[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #42314] Matrix transpose multiply trick - perf
From: |
Ceral Paquet |
Subject: |
[Octave-bug-tracker] [bug #42314] Matrix transpose multiply trick - performance bug? |
Date: |
Wed, 07 May 2014 20:59:32 +0000 |
User-agent: |
Mozilla/5.0 (X11; Linux i686; rv:24.0) Gecko/20100101 Firefox/24.0 |
URL:
<http://savannah.gnu.org/bugs/?42314>
Summary: Matrix transpose multiply trick - performance bug?
Project: GNU Octave
Submitted by: octavebugs
Submitted on: Wed 07 May 2014 08:59:30 PM GMT
Category: None
Severity: 3 - Normal
Priority: 5 - Normal
Item Group: None
Status: None
Assigned to: None
Originator Name:
Originator Email:
Open/Closed: Open
Discussion Lock: Any
Release: 3.8.1
Operating System: GNU/Linux
_______________________________________________________
Details:
This used to be an issue on MATLAB - see below.
>> A=randn(4000,3000);
>> b=randn(4000,1);
>> tic; q=A'*b; toc
Elapsed time is 0.697953 seconds.
>> tic; qq=(b'*A)'; toc
Elapsed time is 0.0830581 seconds.
>> norm(q-qq)
ans = 7.6550e-12
Is it a bug? I don't know. But it was fixed in MATLAB to make both equally
fast.
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?42314>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
- [Octave-bug-tracker] [bug #42314] Matrix transpose multiply trick - performance bug?,
Ceral Paquet <=